Optimize histogram iterators
Histogram iterators allocate new objects in the AtHistogram and
AtFloatHistogram methods, which makes calculating rates over long
ranges expensive.
In #13215 we allowed an existing object to be reused
when converting an integer histogram to a float histogram. This commit follows
the same idea and allows injecting an existing object in the AtHistogram and
AtFloatHistogram methods. When the injected value is nil, iterators allocate
new histograms, otherwise they populate and return the injected object.
The commit also adds a CopyTo method to Histogram and FloatHistogram which
is used in the BufferedIterator to overwrite items in the ring instead of making
new copies.
Note that a specialized HPoint pool is needed for all of this to work
(`matrixSelectorHPool`).
---------
Signed-off-by: Filip Petkovski <filip.petkovsky@gmail.com>
Co-authored-by: George Krajcsovits <krajorama@users.noreply.github.com>
Prometheus is hard-coded to use a fanout storage between TSDB and
a remote storage which by default is empty.
This change detects the empty storage and skips merging between
result sets, which would make `Select()` sort results.
Bottom line: we skip a sort unless there really is some remote storage
configured.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
* Drop old inmemory samples
Co-authored-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Avoid copying timeseries when the feature is disabled
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Run gofmt
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Clarify docs
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Add more logging info
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Remove loggers
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* optimize function and add tests
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Simplify filter
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* rename var
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Update help info from metrics
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* use metrics to keep track of drop elements during buildWriteRequest
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* rename var in tests
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* pass time.Now as parameter
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Change buildwriterequest during retries
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Revert "Remove loggers"
This reverts commit 54f91dfcae20488944162335ab4ad8be459df1ab.
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* use log level debug for loggers
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
* Fix linter
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Remove noisy debug-level logs; add 'reason' label to drop metrics
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Remove accidentally committed files
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Propagate logger to buildWriteRequest to log dropped data
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Fix docs comment
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Make drop reason more specific
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Remove unnecessary pass of logger
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Use snake_case for reason label
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
* Fix dropped samples metric
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
---------
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
Signed-off-by: Paschalis Tsilias <tpaschalis@users.noreply.github.com>
Co-authored-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Co-authored-by: Paschalis Tsilias <tpaschalis@users.noreply.github.com>
They are used in multiple repos, so common is a better place for them.
Several packages now don't depend on `model/textparse`, e.g.
`storage/remote`.
Also remove `metadata` struct from `api.go`, since it was identical to
a struct in the `metadata` package.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
* Append created timestamps.
Signed-off-by: Arthur Silva Sens <arthur.sens@coralogix.com>
* Log when created timestamps are ignored
Signed-off-by: Arthur Silva Sens <arthur.sens@coralogix.com>
* Proposed changes to Append CT PR.
Changes:
* Changed textparse Parser interface for consistency and robustness.
* Changed CT interface to be more explicit and handle validation.
* Simplified test, change scrapeManager to allow testability.
* Added TODOs.
Signed-off-by: bwplotka <bwplotka@gmail.com>
* Updates.
Signed-off-by: bwplotka <bwplotka@gmail.com>
* Addressed comments.
Signed-off-by: bwplotka <bwplotka@gmail.com>
* Refactor head_appender test
Signed-off-by: Arthur Silva Sens <arthur.sens@coralogix.com>
* Fix linter issues
Signed-off-by: Arthur Silva Sens <arthur.sens@coralogix.com>
* Use model.Sample in head appender test
Signed-off-by: Arthur Silva Sens <arthur.sens@coralogix.com>
---------
Signed-off-by: Arthur Silva Sens <arthur.sens@coralogix.com>
Signed-off-by: bwplotka <bwplotka@gmail.com>
Co-authored-by: bwplotka <bwplotka@gmail.com>
This commit reduces the memory needed to query native histogram objects
by reusing existing HPoint instances.
Signed-off-by: Filip Petkovski <filip.petkovsky@gmail.com>
The 'ToFloat' method on integer histograms currently allocates new memory
each time it is called.
This commit adds an optional *FloatHistogram parameter that can be used
to reuse span and bucket slices. It is up to the caller to make sure the
input float histogram is not used anymore after the call.
Signed-off-by: Filip Petkovski <filip.petkovsky@gmail.com>
The ChunkReader interface's Chunk() has been changed to ChunkOrIterable().
This is a precursor to OOO native histogram support - with OOO native histograms, the chunks.Meta passed to Chunk() can result in multiple chunks being returned rather than just a single chunk (e.g. if oooMergedChunk has a counter reset in the middle).
To support this, ChunkOrIterable() requires either a single chunk or an iterable to be returned. If an iterable is returned, the caller has the responsibility of converting the samples from the iterable into possibly multiple chunks. The OOOHeadChunkReader now returns an iterable rather than a chunk to prepare for the native histograms case. Also as a beneficial side effect, oooMergedChunk and boundedChunk has been simplified as they only need to implement the Iterable interface now, not the full Chunk interface.
---------
Signed-off-by: Fiona Liao <fiona.y.liao@gmail.com>
Co-authored-by: George Krajcsovits <krajorama@users.noreply.github.com>
Too confusing to have `MetadataList` and `ListMetadata`, etc.
I standardised on the ones which are in an interface.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
I initially didn't copy the otlptranslator/prometheus folder because I
assumed it wouldn't get changes. But it did. So this PR fixes that and
updates the Collector version.
Supersedes: https://github.com/prometheus/prometheus/pull/12809
Signed-off-by: Goutham <gouthamve@gmail.com>
* Fix issue where `chainSampleIterator` can obscure errors
Signed-off-by: Charles Korn <charles.korn@grafana.com>
* Address PR feedback.
Signed-off-by: Charles Korn <charles.korn@grafana.com>
---------
Signed-off-by: Charles Korn <charles.korn@grafana.com>
Two issues are fixed here, that lead to the same problem:
1. If `newSampleRing` is called with an unknown ValueType including
ValueNone, we have initialized the interface buffer (`iBuf`).
However, we would still use a specialized buffer for the first
sample, opportunistically assuming that we might still not
encounter mixed samples and we should go down the more efficient
road.
2. If the `sampleRing` is `reset`, we leave all buffers alone,
including `iBuf`, which is generally fine, but not for `iBuf`, see
below.
In both cases, `iBuf` already contains values, but we will fill one of
the specialized buffers first. Once we then actually encounter mixed
samples, the content of the specialized buffer is copied into `iBuf`
using `append`. That's by itself the right idea because `iBuf` might
be `nil`, and even if not, it might or might not have the right
capacity. However, this approach assumes that `iBuf` is empty, or more
precisely has a length of zero.
This commit makes sure that `iBuf` does not get needlessly initialized
in `newSampleRing` and that it is emptied upon `reset`.
A test case is added to demonstrate both issues above.
Signed-off-by: beorn7 <beorn@grafana.com>
Instead of setting to nil and allocating a new slice every time the
merge is advanced, re-use the previous slice.
This is safe because the `currentSets` member is only used inside member
functions, and explicitly copied in `At()`, the only place it leaves the
struct.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
The new version features a set of test cases that simplify the addition
of new HTTP status codes.
Signed-off-by: William Dumont <william.dumont@grafana.com>
Header name is `Retry-Attempt`, only set when >0.
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
Signed-off-by: Paschalis Tsilias <paschalis.tsilias@grafana.com>
Return annotations (warnings and infos) from PromQL queries
This generalizes the warnings we have already used before (but only for problems with remote read) as "annotations".
Annotations can be warnings or infos (the latter could be false positives). We do not treat them different in the API for now and return them all as "warnings". It would be easy to distinguish them and return infos separately, should that appear useful in the future.
The new annotations are then used to create a lot of warnings or infos during PromQL evaluations. Partially these are things we have wanted for a long time (e.g. inform the user that they have applied `rate` to a metric that doesn't look like a counter), but the new native histograms have created even more needs for those annotations (e.g. if a query tries to aggregate float numbers with histograms).
The annotations added here are not yet complete. A prominent example would be a warning about a range too short for a rate calculation. But such a warnings is more tricky to create with good fidelity and we will tackle it later.
Another TODO is to take annotations into account when evaluating recording rules.
---------
Signed-off-by: Jeanette Tan <jeanette.tan@grafana.com>
promql: Extend testing framework to support native histograms
This includes both the internal testing framework as well as the rules unit test feature of promtool.
This also adds a bunch of basic tests. Many of the code level tests can now be converted to tests within the framework, and more tests can be added easily.
---------
Signed-off-by: Harold Dost <h.dost@criteo.com>
Signed-off-by: Gregor Zeitlinger <gregor.zeitlinger@grafana.com>
Signed-off-by: Stephen Lang <stephen.lang@grafana.com>
Co-authored-by: Harold Dost <h.dost@criteo.com>
Co-authored-by: Stephen Lang <stephen.lang@grafana.com>
Co-authored-by: Gregor Zeitlinger <gregor.zeitlinger@grafana.com>
Add a chunk size limit in bytes
This creates a hard cap for XOR chunks of 1024 bytes.
The limit for histogram chunk is also 1024 bytes, but it is a soft limit as a histogram has a dynamic size, and even a single one could be larger than 1024 bytes.
This also avoids cutting new histogram chunks if the existing chunk has fewer than 10 histograms yet. In that way, we are accepting "jumbo chunks" in order to have at least 10 histograms in a chunk, allowing compression to kick in.
Signed-off-by: Justin Lei <justin.lei@grafana.com>
So far, `ValidateHistogram` would not detect if the count did not
include the count in the zero bucket. This commit fixes the problem
and updates all the tests that have been undetected offenders so far.
Note that this problem would only ever create false negatives, so we
never falsely rejected to store a histogram because of it.
On the other hand, `ValidateFloatHistogram` has been to strict with
the count being at least as large as the sum of the counts in all the
buckets. Float precision issues could create false positives here, see
products of PromQL evaluations, it's actually quite hard to put an
upper limit no the floating point imprecision. Users could produce the
weirdest expressions, maxing out float precision problems. Therefore,
this commit simply removes that particular check from
`ValidateFloatHistogram`.
Signed-off-by: beorn7 <beorn@grafana.com>
* Add OTLP Ingestion endpoint
We copy files from the otel-collector-contrib. See the README in
`storage/remote/otlptranslator/README.md`.
This supersedes: https://github.com/prometheus/prometheus/pull/11965
Signed-off-by: gouthamve <gouthamve@gmail.com>
* Return a 200 OK
It is what the OTEL Golang SDK expect :(
https://github.com/open-telemetry/opentelemetry-go/issues/4363
Signed-off-by: Goutham <gouthamve@gmail.com>
---------
Signed-off-by: gouthamve <gouthamve@gmail.com>
Signed-off-by: Goutham <gouthamve@gmail.com>
* WIP implement WAL watcher reading via notifications over a channel from
the TSDB code
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* Notify via head appenders Commit (finished all WAL logging) rather than
on each WAL Log call
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* Fix misspelled Notify plus add a metric for dropped Write notifications
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* Update tests to handle new notification pattern
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* this test maybe needs more time on windows?
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* does this test need more time on windows as well?
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* read timeout is already a time.Duration
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* remove mistakenly commited benchmark data files
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* address some review feedback
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* fix missed changes from previous commit
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* Fix issues from wrapper function
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* try fixing race condition in test by allowing tests to overwrite the
read ticker timeout instead of calling the Notify function
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* fix linting
Signed-off-by: Callum Styan <callumstyan@gmail.com>
---------
Signed-off-by: Callum Styan <callumstyan@gmail.com>
The storage.ChunkSeries iterator assumes that a histogram sample can always be
appended to the currently open chunk. This is not the case when there is a counter reset,
or when appending a stale sample to a chunk with non-stale samples. In addition, the open chunk sometimes
needs to be recoded before a sample can be appended.
This commit addresses the issue by implementing a RecodingAppender which can recode incoming
samples in a transparent way. It also detects cases when a sample cannot be appended at all and
returns `false` so that the caller can open a new chunk.
Signed-off-by: Filip Petkovski <filip.petkovsky@gmail.com>
Signed-off-by: György Krajcsovits <gyorgy.krajcsovits@grafana.com>
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
Co-authored-by: György Krajcsovits <gyorgy.krajcsovits@grafana.com>
Co-authored-by: Ganesh Vernekar <ganeshvern@gmail.com>
Wiser coders than myself have come to the conclusion that a `switch`
statement is almost always superior to a statement that includes any
`else if`.
The exceptions that I have found in our codebase are just these two:
* The `if else` is followed by an additional statement before the next
condition (separated by a `;`).
* The whole thing is within a `for` loop and `break` statements are
used. In this case, using `switch` would require tagging the `for`
loop, which probably tips the balance.
Why are `switch` statements more readable?
For one, fewer curly braces. But more importantly, the conditions all
have the same alignment, so the whole thing follows the natural flow
of going down a list of conditions. With `else if`, in contrast, all
conditions but the first are "hidden" behind `} else if `, harder to
spot and (for no good reason) presented differently from the first
condition.
I'm sure the aforemention wise coders can list even more reasons.
In any case, I like it so much that I have found myself recommending
it in code reviews. I would like to make it a habit in our code base,
without making it a hard requirement that we would test on the CI. But
for that, there has to be a role model, so this commit eliminates all
`if else` occurrences, unless it is autogenerated code or fits one of
the exceptions above.
Signed-off-by: beorn7 <beorn@grafana.com>
We haven't updated golint-ci in our CI yet, but this commit prepares
for that.
There are a lot of new warnings, and it is mostly because the "revive"
linter got updated. I agree with most of the new warnings, mostly
around not naming unused function parameters (although it is justified
in some cases for documentation purposes – while things like mocks are
a good example where not naming the parameter is clearer).
I'm pretty upset about the "empty block" warning to include `for`
loops. It's such a common pattern to do something in the head of the
`for` loop and then have an empty block. There is still an open issue
about this: https://github.com/mgechev/revive/issues/810 I have
disabled "revive" altogether in files where empty blocks are used
excessively, and I have made the effort to add individual
`// nolint:revive` where empty blocks are used just once or twice.
It's borderline noisy, though, but let's go with it for now.
I should mention that none of the "empty block" warnings for `for`
loop bodies were legitimate.
Signed-off-by: beorn7 <beorn@grafana.com>
Signed-off-by: Sebastian Rabenhorst <sebastian.rabenhorst@shopify.com>
Fixed sampleRingIterator for mixed histograms
Signed-off-by: Sebastian Rabenhorst <sebastian.rabenhorst@shopify.com>
Fixed lint
This commit is doing what I would have expected that Go generics do
for me. However, the PromQL benchmarks show a significant runtime and
allocation increase with `genericAdd`, so this replaces it with
hand-coded non-generic versions of it.
Signed-off-by: beorn7 <beorn@grafana.com>
In the past, every sample value was a float, so it was fine to call a
variable holding such a float "value" or "sample". With native
histograms, a sample might have a histogram value. And a histogram
value is still a value. Calling a float value just "value" or "sample"
or "V" is therefore misleading. Over the last few commits, I already
renamed many variables, but this cleans up a few more places where the
changes are more invasive.
Note that we do not to attempt naming in the JSON APIs or in the
protobufs. That would be quite a disruption. However, internally, we
can call variables as we want, and we should go with the option of
avoiding misunderstandings.
Signed-off-by: beorn7 <beorn@grafana.com>
This utilizes the fact that most sampleRings will only contain samples
of one type. In this case, the generic interface is circumvented, and
a bespoke buffer for the one actually occurring sample type is
used. Should a sampleRing receive a sample of a different kind later,
it will transparently switch to the generic behavior.
Signed-off-by: beorn7 <beorn@grafana.com>
Previously, we had one “polymorphous” `sample` type in the `storage`
package. This commit breaks it up into `fSample`, `hSample`, and
`fhSample`, each still implementing the `tsdbutil.Sample` interface.
This reduces allocations in `sampleRing.Add` but inflicts the penalty
of the interface wrapper, which makes things worse in total.
This commit therefore just demonstrates the step taken. The next
commit will tackle the interface overhead problem.
Signed-off-by: beorn7 <beorn@grafana.com>
It took a `Labels` where the memory could be re-used, but in practice
this hardly ever benefitted. Especially after converting `relabel.Process`
to `relabel.ProcessBuilder`.
Comparing the parameter to `nil` was a bug; `EmptyLabels` is not `nil`
so the slice was reallocated multiple times by `append`.
Lastly `Builder.Labels()` now estimates that the final size will depend
on labels added and deleted.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
During remote write, we call url.String() twice:
- to add the Endpoint() to the span
- to actually know where whe should send the request
This value does not change over time, and it's not really that
lightweight to calculate. I wrote this simple benchmark:
func BenchmarkURLString(b *testing.B) {
u, err := url.Parse("https://remote.write.com/api/v1")
require.NoError(b, err)
b.Run("string", func(b *testing.B) {
count := 0
for i := 0; i < b.N; i++ {
count += len(u.String())
}
})
}
And the results are ~200ns/op, 80B/op, 3 allocs/op.
Yes, we're going to go to the network here, which is a huge amount of
resources compared to this, but still, on agents that send 500 requests
per second, that is 1500 wasteful allocations per second.
Signed-off-by: Oleg Zaytsev <mail@olegzaytsev.com>
In storage/remote, try converting to RecoverableError using errors.As,
instead of through direct casting.
Signed-off-by: Arve Knudsen <arve.knudsen@gmail.com>
As far as I understand it, we'd never expect to receive a nil span,
and remote.spansProtoToSpans would panic if we received a nil span.
Marking the fields as non-nullable also means the generated Golang
code doesn't use pointers for these fields, reducing allocations.
Signed-off-by: Charles Korn <charles.korn@grafana.com>
This is a bit more conservative than we could be. As long as a chunk
isn't the first in a block, we can be pretty sure that the previous
chunk won't disappear. However, the incremental gain of returning
NotCounterReset in these cases is probably very small and might not be
worth the code complications.
Wwith this, we now also pay attention to an explicitly set counter
reset during ingestion. While the case doesn't show up in practice
yet, there could be scenarios where the metric source knows there was
a counter reset even if it might not be visible from the values in the
histogram. It is also useful for testing.
Signed-off-by: beorn7 <beorn@grafana.com>
This is an optimization on the existing append in OOOChunk.
What we've been doing so far is find the place inside the out-of-order
slice where the new sample should go in and then place it there and move
any samples to the right if necessary. This is OK but requires a binary
search every time the slice is bigger than 0.
The optimization is opinionated and suggests that although out-of-order
samples can be out-of-order amongst themselves they'll probably be in
order thus we can probably optimistically append at the end and if not
do the binary search.
OOOChunks are capped to 30 samples by default so this is a small
optimization but everything adds up, specially if you handle many active
timeseries with out-of-order samples.
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
Signed-off-by: Jesus Vazquez <jesusvazquez@users.noreply.github.com>
Co-authored-by: Ganesh Vernekar <ganeshvern@gmail.com>
* adapt code.go and write_handler.go to support float histograms
* adapt watcher.go to support float histograms
* wip adapt queue_manager.go to support float histograms
* address comments for metrics in queue_manager.go
* set test cases for queue manager
* use same counts for histograms and float histograms
* refactor createHistograms tests
* fix float histograms ref in watcher_test.go
* address PR comments
Signed-off-by: Marc Tuduri <marctc@protonmail.com>
In most cases, there is no sample at `maxt`, so `PeekBack` has to be
used. So far, `PeekBack` did not return a float histogram, and we
disregarded even any returned normal histogram. This fixes both, and
also tweaks the unit test to discover the problem (by using an earlier
timestamp than "now" for the samples in the TSDB).
Signed-off-by: beorn7 <beorn@grafana.com>
Extends Appender.AppendHistogram function to accept the FloatHistogram. TSDB supports appending, querying, WAL replay, for this new type of histogram.
Signed-off-by: Marc Tudurí <marctc@protonmail.com>
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
Co-authored-by: Ganesh Vernekar <ganeshvern@gmail.com>
`QueueManager.externalLabels` becomes a slice rather than a `Labels` so
we can index into it when doing the merge operation.
Note we avoid calling `Labels.Len()` in `labelProtosToLabels()`.
It isn't necessary - `append()` will enlarge the buffer and we're
expecting to re-use it many times.
Also, we now validate protobuf input before converting to Labels.
This way we can detect errors first, and we don't place unnecessary
requirements on the Labels structure.
Re-do seriesFilter using labels.Builder (albeit N^2).
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
This is required to preserve the interface property of SeriesSet that says
"At returns full series. Returned series should be iterable even after Next is called."
Signed-off-by: sniper91 <kevinzhao91@outlook.com>
Re-use previous memory if it is already of the correct type.
In `NewListSeries` we hoist the conversion to an interface value out
so it only allocates once.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
Patterned after `Chunk.Iterator()`: pass the old iterator in so it
can be re-used to avoid allocating a new object.
(This commit does not do any re-use; it is just changing all the method
signatures so re-use is possible in later commits.)
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
The SeriesSets to be merged must be created each time round the loop,
otherwise the benchmark is not doing any real work.
Don't call ExpandSeries, because it spends most of its time allocating
a memory buffer to hold the result, which we don't look at.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
Fix up merge test again
errors.Unwrap() actually dangerously returns nil if the error does not have an
Unwrap() method, which is the case in at least one of these places where I
noticed that no error was being logged at all when it should have.
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* remote/read_handler: pool input to Marshal()
Use a sync.Pool to reuse byte slices between calls to Marshal() in the
remote read handler.
Signed-off-by: Giedrius Statkevičius <giedrius.statkevicius@vinted.com>
* remote: add microbenchmark for remote read handler
Signed-off-by: Giedrius Statkevičius <giedrius.statkevicius@vinted.com>
Signed-off-by: Giedrius Statkevičius <giedrius.statkevicius@vinted.com>
The wlog.WL type can now be used to create a Write Ahead Log or a Write
Behind Log.
Before the prefix for wbl metrics was
'prometheus_tsdb_out_of_order_wal_' and has been replaced with
'prometheus_tsdb_out_of_order_wbl_'.
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
Signed-off-by: Jesus Vazquez <jesusvazquez@users.noreply.github.com>
Co-authored-by: Ganesh Vernekar <15064823+codesome@users.noreply.github.com>
Where the code was multiplying bytes by number of operations, this
resulted in absurdly high throughput numbers.
Also, in `BenchmarkParse()`, don't run the `expfmt` case twice.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
Use new experimental package `golang.org/x/exp/slices`.
slices.Sort works on values that are directly comparable, like ints,
so avoids the overhad of an interface call to `.Less()`.
Left tests unchanged, because they don't need the speed and it may be
a cross-check that slices.Sort gives the same answer.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
* Introduce out-of-order TSDB support
This implementation is based on this design doc:
https://docs.google.com/document/d/1Kppm7qL9C-BJB1j6yb6-9ObG3AbdZnFUBYPNNWwDBYM/edit?usp=sharing
This commit adds support to accept out-of-order ("OOO") sample into the TSDB
up to a configurable time allowance. If OOO is enabled, overlapping querying
are automatically enabled.
Most of the additions have been borrowed from
https://github.com/grafana/mimir-prometheus/
Here is the list ist of the original commits cherry picked
from mimir-prometheus into this branch:
- 4b2198d7ec
- 2836e5513f
- 00b379c3a5
- ff0dc75758
- a632c73352
- c6f3d4ab33
- 5e8406a1d4
- abde1e0ba1
- e70e769889
- df59320886
Co-authored-by: Jesus Vazquez <jesus.vazquez@grafana.com>
Co-authored-by: Ganesh Vernekar <ganeshvern@gmail.com>
Co-authored-by: Dieter Plaetinck <dieter@grafana.com>
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* gofumpt files
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Add license header to missing files
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Fix OOO tests due to existing chunk disk mapper implementation
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Fix truncate int overflow
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Add Sync method to the WAL and update tests
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* remove useless sync
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Update minOOOTime after truncating Head
* Update minOOOTime after truncating Head
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Fix lint
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Add a unit test
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Load OutOfOrderTimeWindow only once per appender
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Fix OOO Head LabelValues and PostingsForMatchers
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Fix replay of OOO mmap chunks
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Remove unnecessary err check
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Prevent panic with ApplyConfig
Signed-off-by: Ganesh Vernekar 15064823+codesome@users.noreply.github.com
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Run OOO compaction after restart if there is OOO data from WBL
Signed-off-by: Ganesh Vernekar 15064823+codesome@users.noreply.github.com
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Apply Bartek's suggestions
Co-authored-by: Bartlomiej Plotka <bwplotka@gmail.com>
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Refactor OOO compaction
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Address comments and TODOs
- Added a comment explaining why we need the allow overlapping
compaction toggle
- Clarified TSDBConfig OutOfOrderTimeWindow doc
- Added an owner to all the TODOs in the code
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Run go format
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Fix remaining review comments
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Fix tests
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Change wbl reference when truncating ooo in TestHeadMinOOOTimeUpdate
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
* Fix TestWBLAndMmapReplay test failure on windows
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Address most of the feedback
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Refactor the block meta for out of order
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Fix windows error
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Fix review comments
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
Signed-off-by: Jesus Vazquez <jesus.vazquez@grafana.com>
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
Signed-off-by: Ganesh Vernekar 15064823+codesome@users.noreply.github.com
Co-authored-by: Ganesh Vernekar <15064823+codesome@users.noreply.github.com>
Co-authored-by: Ganesh Vernekar <ganeshvern@gmail.com>
Co-authored-by: Dieter Plaetinck <dieter@grafana.com>
Co-authored-by: Oleg Zaytsev <mail@olegzaytsev.com>
Co-authored-by: Bartlomiej Plotka <bwplotka@gmail.com>
* Add runtime config to control native histogram ingestion
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
* Make the config into a CLI flag
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
Signed-off-by: Ganesh Vernekar <ganeshvern@gmail.com>
And a few cases of `EmptyLabels()`.
Replacing code which assumes the internal structure of `Labels`.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
* Add histogram validation
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Correct negative offset validation
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Address review comments
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Validation benchmark
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Add more checks
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Attempt to fix tests
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Fix stuff
Signed-off-by: Levi Harrison <git@leviharrison.dev>
* Append metadata to the WAL
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Remove extra whitespace; Reword some docstrings and comments
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Use RLock() for hasNewMetadata check
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Use single byte for metric type in RefMetadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Update proposed WAL format for single-byte type metadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Implementa MetadataAppender interface for the Agent
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Address first round of review comments
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Amend description of metadata in wal.md
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Correct key used to retrieve metadata from cache
When we're setting metadata entries in the scrapeCace, we're using the
p.Help(), p.Unit(), p.Type() helpers, which retrieve the series name and
use it as the cache key. When checking for cache entries though, we used
p.Series() as the key, which included the metric name _with_ its labels.
That meant that we were never actually hitting the cache. We're fixing
this by utiling the __name__ internal label for correctly getting the
cache entries after they've been set by setHelp(), setType() or
setUnit().
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Put feature behind a feature flag
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix AppendMetadata docstring
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Reorder WAL format document
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Change error message of AppendMetadata; Fix access of s.meta in AppendMetadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Reuse temporary buffer in Metadata encoder
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Only keep latest metadata for each refID during checkpointing
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix test that's referencing decoding metadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Avoid creating metadata block if no new metadata are present
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Add tests for corrupt metadata block and relevant record type
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix CR comments
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Extract logic about changing metadata in an anonymous function
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Implement new proposed WAL format and amend relevant tests
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Use 'const' for metadata field names
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Apply metadata to head memSeries in Commit, not in AppendMetadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Add docstring and rename extracted helper in scrape.go
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Add tests for tsdb-related cases
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix linter issues vol1
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix linter issues vol2
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix Windows test by closing WAL reader files
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Use switch instead of two if statements in metadata decoding
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix review comments around TestMetadata* tests
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Add code for replaying WAL; test correctness of in-memory data after a replay
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Remove scrape-loop related code from PR
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Address first round of comments
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Simplify tests by sorting slices before comparison
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix test to use separate transactions
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Empty out buffer and record slices after encoding latest metadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix linting issue
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Update calculation for DroppedMetadata metric
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Rename MetadataAppender interface and AppendMetadata method to MetadataUpdater/UpdateMetadata
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Reuse buffer when encoding latest metadata for each series
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Fix review comments; Check all returned error values using two helpers
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Simplify use of helpers
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Satisfy linter
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
Note: This is deliberately an incompatible change. Since we have never
used histograms in remote read/write yet, there is no point in keeping
compatibility. This _is_, however, compatible to the state in the main
branch.
This commit flattens the bucket message into top-level fields. This
has the disadvantage of now having two triples of fields prefixed with
`negative_...` or `positive_...`. However, with this structure, we
save one tag on the wire. And, perhaps more importantly, we mirror the
structure of the `histogram.Histogram` Go type.
This commit also adjusts `repeated` fields to use names in the plural
form, as it is also the case for the fields that already existed.
This also adds a doc comment to `HistogramProtoToHistogram` and
changes its return type to a pointer (which is more convenient and
probably more efficient).
Signed-off-by: beorn7 <beorn@grafana.com>
* Removing global state modification on unit tests (fix#10033#10034)
The config.DefaultRemoteReadConfig and config.DefaultRemoteWriteConfig
instances hold global state. Unit tests were changing their url.URL reference
globally causing false positives when tests were ran through package.
Two helper functions were created to copy those global values instead of changing
them in place to fix null point when running unit tests by method instead of
by package.
Signed-off-by: Leonardo Zamariola <leonardo.zamariola@gmail.com>
* Fixing pull request suggestions
Copying by value from default config
Signed-off-by: Leonardo Zamariola <leonardo.zamariola@gmail.com>
* Rename walDir parameter to dir
Signed-off-by: Matej Gera <matejgera@gmail.com>
* Improve NewQueueManager comment
Signed-off-by: Matej Gera <matejgera@gmail.com>
"Labels is a sorted set of labels. Order has to be guaranteed upon
instantiation." says the comment, so fix all the tests that break this
rule.
For `BenchmarkLabelValuesWithMatchers()` and
`BenchmarkHeadLabelValuesWithMatchers()` the amount of work done changes
significantly if you put the labels in order, because all series refs
get neatly partitioned by the `tens` label, so I renamed the labels
to maintain the previous behaviour.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
* refactor: move from io/ioutil to io and os packages
* use fs.DirEntry instead of os.FileInfo after os.ReadDir
Signed-off-by: MOREL Matthieu <matthieu.morel@cnp.fr>
If FlushAndShutdown is called with a full batchQueue, and then Batch is
called rather than the normal path of reading from a queue a deadlock
might be encountered. Rather than having FlushAndShutdown having
blocking code while holding a lock retry sending the batch every second.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
Do not block when trying to write a batch to the queue. This can cause
appends to lock forever if the only thing reading from the queue needs
the mutex to write. Instead, if batchQueue is full pop the sample that
was just added from the partial batch and return false. The code doing
the appending already handles retries with backoff.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
* Run gofumpt on all files
Getting golangci-lint errors when building on my laptop, possibly because I have newer version of gofumpt then what it was formatted with.
Run gofumpt -w -extra on all files as it will be needed in the future anyway.
* Update golangci-lint to v1.44.2
v1.44.0 upgraded gofumpt so bumping version in CI will help keep formatting correct for everyone
* Address golangci-lint error
Getting 'error-strings: error strings should not be capitalized or end with punctuation or a newline' from revive here.
Drop new line.
Signed-off-by: Łukasz Mierzwa <l.mierzwa@gmail.com>
If a queue is stopped and one of its shards happens to hit the
batch_send_deadline at the same time a deadlock can occur where stop
holds the mutex and will not release it until the send is finished, but
the send needs the mutex to retrieve the most recent batch. This is
fixed by using a second mutex just for writing.
In addition, the test I wrote exposed a case where during shutdown a
batch could be sent twice due to concurrent calls to queue.Batch() and
queue.FlushAndShutdown(). Protect these with a mutex as well.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
Previously we would reject an increase from 2 to 2.5 as being
within 30%; by rounding up first we see this as an increase from 2 to 3.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
Change the coefficient from 1% to 5%, so instead of targetting to clear
the backlog in 100s we target 20s.
Update unit test to reflect the new behaviour.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
We have an alert that fires when prometheus_remote_storage_highest_timestamp_in_seconds - prometheus_remote_storage_queue_highest_sent_timestamp_seconds
becomes too high. But we have an agent that fires this when the remote "rate-limits" the user.
This is because prometheus_remote_storage_queue_highest_sent_timestamp_seconds doesn't get updated
when the remote sends a 429.
I think we should update the metrics, and the change I made makes sense. Because if the requests fails
because of connectivity issues, etc. we will never exit the `sendWriteRequestWithBackoff` function. It only
exits the function when there is a non-recoverable error, like a bad status code, and in that case, I think
the metric needs to be updated.
Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
* storage: expose bug in iterators #10027
Signed-off-by: beorn7 <beorn@grafana.com>
* storage: fix bug #10027 in iterators' Seek method
Signed-off-by: beorn7 <beorn@grafana.com>
* Append reporting metrics without limit
If reporting metrics fails due to reaching the limit, this makes the
target appear as UP in the UI, but the metrics are missing.
This commit bypasses that limit for report metrics.
Signed-off-by: Julien Pivotto <roidelapluie@inuits.eu>
* Remove check against cfg so interval/ timeout are always set (#10023) (#10031)
Signed-off-by: Nicholas Blott <blottn@tcd.ie>
Co-authored-by: Nicholas Blott <blottn@tcd.ie>
* Cut v2.32.1
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* Apply suggestions from code review
Signed-off-by: Julius Volz <julius.volz@gmail.com>
Co-authored-by: Levi Harrison <git@leviharrison.dev>
Co-authored-by: Julien Pivotto <roidelapluie@inuits.eu>
Co-authored-by: Nicholas Blott <blottn@tcd.ie>
Co-authored-by: Julius Volz <julius.volz@gmail.com>
Co-authored-by: Levi Harrison <git@leviharrison.dev>
Previously BenchmarkSampleDelivery spent a lot of effort checking each
sample had arrived, so was largely showing the performance of test-only
code.
Increase the number of shards to be more realistic for a large
workload.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
Right now the values for enqueuedSamples and enqueuedExemplars is never
subtracted leading to inflated values for failedSamples/failedExemplars
when a hard shutdown of a shard occurs.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
Channels can cause bottlenecks and tons of context switches when reading
hundreds of thousands of samples per second from a single queue.
Instead, pre-batch the samples to amortize the cost of the concurrency
overhead.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
`BufferedSeriesIterator` and `MemoizedSeriesIterator` use a method
called `Values` for exactly the purpose for which all other iterators
of the same kind use a method called `At`. That alone is confusing,
but on top of that, the `Values` method only returns a single sample,
not multiple values. I assume the naming has historical reasons. This
commit makes it more consistent. It is now easier to read, and now
`BufferedSeriesIterator` and `MemoizedSeriesIterator` implement
`chunkenc.Iterator` like many other iterators, too.
Signed-off-by: beorn7 <beorn@grafana.com>
There was a subtle and nasty bug in listSeriesIterator.Seek.
In addition, the Seek call is defined to be a no-op if the current
position of the iterator is already pointing to a suitable
sample. This commit adds fast paths for this case to several
potentially expensive Seek calls.
Another bug was in concreteSeriesIterator.Seek. It always searched the
whole series and not from the current position of the iterator.
Signed-off-by: beorn7 <beorn@grafana.com>
- Pick At... method via return value of Next/Seek.
- Do not clobber returned buckets.
- Add partial FloatHistogram suppert.
Note that the promql package is now _only_ dealing with
FloatHistograms, following the idea that PromQL only knows float
values.
As a byproduct, I have removed the histogramSeries metric. In my
understanding, series can have both float and histogram samples, so
that metric doesn't make sense anymore.
As another byproduct, I have converged the sampleBuf and the
histogramSampleBuf in memSeries into one. The sample type stored in
the sampleBuf has been extended to also contain histograms even before
this commit.
Signed-off-by: beorn7 <beorn@grafana.com>
