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>
* Add a test with variable samples rate append
This test overflows the chunk created in memseries, and the total amount
of samples in the (only) mmapped chunk is 29, instead of the 65565
appended ones.
Signed-off-by: Oleg Zaytsev <mail@olegzaytsev.com>
* Cut new chunk when rate prediction was wrong
When appending samples at a slow rate, and then appending at a higher
rate, the prediction we made to cut a new chunk is no longer valid.
Sometimes this can even cause an overflow in the chunk, if more samples
than uint16 can hold are appended.
Signed-off-by: Oleg Zaytsev <mail@olegzaytsev.com>
* Improve comment on 2*samplesPerChunk
Signed-off-by: Oleg Zaytsev <mail@olegzaytsev.com>
* Assert that all chunks have less than 240 samples
Also, trigger new chunk at 240, not at more than 240
Signed-off-by: Oleg Zaytsev <mail@olegzaytsev.com>
* tsdb/agent: Ignore duplicate exemplars
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Make each exemplar unique in TestCommit
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Re-Trigger CI for Windows and UI-related steps
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Change test comment to properly re-trigger pipeline
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
* Defer Close() calls for test agent and segment reader
Signed-off-by: Paschalis Tsilias <paschalist0@gmail.com>
OTel Collector prints the following error when a target disappears:
```
2022-04-13T14:20:24.932-0400 warn scrape/scrape.go:1408 Stale append failed {"kind": "receiver", "name": "prometheus", "scrape_pool": "beep-boop", "target": "http://localhost:9090/metrics", "error": "transaction aborted"}
```
This `transaction aborted` error is returned by the custom appender that is
used by the collector when the context of the appender is cancelled:
b7bf11174e/receiver/prometheusreceiver/internal/otlp_transaction.go (L81-L82)
We call `endOfRunStaleness` after `sl.stop()` which cancels `sl.ctx`.
The other `.Appender()` calls use `parentCtx` for the same reason.
This hasn't come up so far because Prometheus' Appender implementation just
ignores the context passed.
Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
Storing the scrape cache and the target (which also contains that cache)
is apparently causing hige memory increase. I think me might not control
the lifespan of the context enough, therefore old objects keep living in
memory for longer than needed.
Let's unblock the release and look for an alternative so that downstream
consumers can get access to that data.
Signed-off-by: Julien Pivotto <roidelapluie@o11y.eu>
* tsdb/agent: port grafana/agent#676grafana/agent#676 fixed an issue where a loading a WAL with multiple
segments may result in ref ID collision. The starting ref ID for new
series should be set to the highest ref ID across all series records
from all WAL segments.
This fixes an issue where the starting ref ID was incorrectly set to the
highest ref ID found in the newest segment, which may not have any ref
IDs at all if no series records have been appended to it yet.
Signed-off-by: Robert Fratto <robertfratto@gmail.com>
* tsdb/agent: update terminology (s/ref ID/nextRef)
Signed-off-by: Robert Fratto <robertfratto@gmail.com>
This will make UI static files part of the release artefacts, for
consumption by downstream distributions.
Signed-off-by: Julien Pivotto <roidelapluie@inuits.eu>
* Proper support for go modules.
This pull requests makes Prometheus go-mod compatible.
The general idea is to release the Prometheus libraries as v0.x
releases, next to the v2.x tags used by end users.
This is done by mirroring Prometheus 2.x tags with Prometheus 0.x tags.
When v2.X.0 is released, we would release v0.X.0.
Pre-go mod versions are retracted from go.mod. This is not nice but
should work.
Only v2.x tags will be built and released by CI. v0.x.x tags would just
be normal tags in the repo, not promoted as releases.
Signed-off-by: Julien Pivotto <roidelapluie@inuits.eu>
The tarball should be compressed and only use uncompressed assets. In
that way, uit can be used by downstream distros easily. they can either
download the assets and serve it as is or use the compress script to
have it in the binary.
Signed-off-by: Julien Pivotto <roidelapluie@inuits.eu>
Some downstream distros fail to compile Prometheus UI (Debian, NixOS).
This is an attempt to store compiled UI in circleci for them to consume
it. Currently we do not expose this outside of circleci.
Signed-off-by: Julien Pivotto <roidelapluie@inuits.eu>
Co-authored-by: Julien Pivotto <roidelapluie@o11y.eu>
* Add a --lint flag to the promtool check rules and check config commands
Checking rules with promtool emits warnings in the case of duplicate rules.
These warnings do not result in a non-zero exit code and are difficult to
spot in CI environments. Additionally, checking for duplicates is closer
to a lint check rather than a syntax check.
This commit adds a --lint flag to commands which include checking rules.
The flag can be used to enable or disable certain linting options
and cause the execution to return a non-zero exit code in case
those options are not met.
Signed-off-by: fpetkovski <filip.petkovsky@gmail.com>
* Exit with status 3 on lint error
Signed-off-by: fpetkovski <filip.petkovsky@gmail.com>
* create lezer-promql module + move codemirror to a pure esm module + unified dependencies
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* ignore test utils file and remove the type "module" in package.json
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* use jest to run the lezer-promql test
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* give an automatic way to update the ui dependencies
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* update all dependencies using make update-npm-deps
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* fix react-app test
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* remove generated file
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* remove unnecessary backslash in script
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* fix reviews
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* rewording
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* use npx to run lezer-generator
Signed-off-by: Augustin Husson <husson.augustin@gmail.com>
* tsdb: avoid slice-to-interface allocation in EnsureOrder
This is pulling the `seriesRefSlice` out of the loop, so the compiler
doesn't allocate a new one on the heap every time.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
* tsdb: use pointer type in Pool for EnsureOrder
As noted by staticcheck, Pool prefers the objects in the pool to have
pointer type. This is a little more fiddly to code, but avoids
allocation of a wrapper object every time a slice is put into the pool.
Removed a comment that said fixing this has a performance penalty: not
borne out by benchmarks.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
