The current separation between lexer and parser is a bit fuzzy when it
comes to operators, aggregators and other keywords. The lexer already
tries to determine the type of a token, even though that type might
change depending on the context.
This led to the problematic behavior that no tokens known to the lexer
could be used as label names, including operators (and, by, ...),
aggregators (count, quantile, ...) or other keywords (for, offset, ...).
This change additionally checks whether an identifier is one of these
types. We might want to check whether the specific item identification
should be moved from the lexer to the parser.
tl;dr: This is not a fundamental solution to the indexing problem
(like tindex is) but it at least avoids utilizing the intersection
problem to the greatest possible amount.
In more detail:
Imagine the following query:
nicely:aggregating:rule{job="foo",env="prod"}
While it uses a nicely aggregating recording rule (which might have a
very low cardinality), Prometheus still intersects the low number of
fingerprints for `{__name__="nicely:aggregating:rule"}` with the many
thousands of fingerprints matching `{job="foo"}` and with the millions
of fingerprints matching `{env="prod"}`. This totally innocuous query
is dead slow if the Prometheus server has a lot of time series with
the `{env="prod"}` label. Ironically, if you make the query more
complicated, it becomes blazingly fast:
nicely:aggregating:rule{job=~"foo",env=~"prod"}
Why so? Because Prometheus only intersects with non-Equal matchers if
there are no Equal matchers. That's good in this case because it
retrieves the few fingerprints for
`{__name__="nicely:aggregating:rule"}` and then starts right ahead to
retrieve the metric for those FPs and checking individually if they
match the other matchers.
This change is generalizing the idea of when to stop intersecting FPs
and go into "retrieve metrics and check them individually against
remaining matchers" mode:
- First, sort all matchers by "expected cardinality". Matchers
matching the empty string are always worst (and never used for
intersections). Equal matchers are in general consider best, but by
using some crude heuristics, we declare some better than others
(instance labels or anything that looks like a recording rule).
- Then go through the matchers until we hit a threshold of remaining
FPs in the intersection. This threshold is higher if we are already
in the non-Equal matcher area as intersection is even more expensive
here.
- Once the threshold has been reached (or we have run out of matchers
that do not match the empty string), start with "retrieve metrics
and check them individually against remaining matchers".
A beefy server at SoundCloud was spending 67% of its CPU time in index
lookups (fingerprintsForLabelPairs), serving mostly a dashboard that
is exclusively built with recording rules. With this change, it spends
only 35% in fingerprintsForLabelPairs. The CPU usage dropped from 26
cores to 18 cores. The median latency for query_range dropped from 14s
to 50ms(!). As expected, higher percentile latency didn't improve that
much because the new approach is _occasionally_ running into the worst
case while the old one was _systematically_ doing so. The 99th
percentile latency is now about as high as the median before (14s)
while it was almost twice as high before (26s).
This offers new semantics in allowing on() for matching
two single-element vectors with no known common labels.
Previosuly this was often done using on(dummy).
This also allows making it explict that you meant
to do an aggregation without labels via by().
Fixes#1597.
The labels listed in the group_ modifier will be copied from the one
side to the many side. It will be valid to specify no labels.
This is intended to replace the existing ON/GROUP_* support.,
The `unless` set operator can be used to return all vector elements from
the LHS which do not match the elements on the RHS. A use case is to
return all metrics for nodes which do not have a specific role:
node_load1 unless on(instance) chef_role{role="app"}
This has the advantage that the user doesn't need
to list all labels they want to keep (as with "by")
but without having to worry about inconsistent labels
as when there's only one time series (as with "keeping_common").
Almost all aggregation should use this rather than the existing
two options as it's much less error prone and easier to maintain
due to not having to always add in "job" plus whatever other common
job-level labels you have like "region".
The documentation speaks about range vectors and range vector selectors.
This change does not fix all issues, we might still expose the term
"Matrix" in error messages using %T.
This change is breaking, use the 'bool' modifier for such comprisons.
After this change all comparisons without 'bool' will filter, and all
comparisons with 'bool' will return 0/1. This makes the language more
consistent and orthogonal, and ultimately easier to learn and use.
If we ever figure out sane semantics for filtering scalar/scalar
comparisons we can add them in, which will most likely come out of how
the new vector() function is used.
This adapts some functionality from the Go standard library for string
literal lexing and unquoting/unescaping.
The following string types are now supported:
Double- or single-quoted strings:
These support all escape sequences that Go supports in double-quoted
string literals. The difference is that Prometheus also has
single-quoted strings (instead of single-quoted runes in Go). Raw
newlines are not allowed.
Backtick-quoted raw strings:
Strings quoted in backticks are treated as raw strings just like in Go
and may contain raw newlines and other special characters directly.
Fixes https://github.com/prometheus/prometheus/issues/1122
Fixes https://github.com/prometheus/prometheus/issues/1121
When doing comparison operations on vectors, filtering
sometimes gets in the way and you have to go to a fair bit of
effort to workaround it in order to always return a result.
The 'bool' modifier instead of filtering returns 0/1 depending
on the result of the compairson.
This is also a prerequisite to removing plain scalar/scalar comparisons,
as it maintains the current behaviour under a new syntax.
`keep_common` is more in line with the function name
`drop_common_labels()` terminology-wise, and also more in line with
`group_left`/`group_right` (no `...ing` verb suffix).
We could also go the full way and call it `keep_common_labels`. That
would have the benefit of being even more consistent with the function
`drop_common_labels()` and would be more explanatory, but it also seems
quite long.
The promql_test checks failure of various bad syntaxed queries.
Those are moved into the parser tests as the new testing language
only deals with valid queries.
This commit adds parsing of time series description to the exisiting
query language parser. Time series descriptions are defined by a
metric followed by a sequence of values.
This copies the evaluation logic from the current rules/ package.
The new engine handles the execution process from query string to final result.
It provides query timeout and cancellation and general flexibility for
future changes.
functions.go: Add evaluation implementation. Slight changes to in/out data but
not to the processing logic.
quantile.go: No changes.
analyzer.go: No changes.
engine.go: Actually new part. Mainly consists of evaluation methods
which were not changed.
setup_test.go: Copy of rules/helpers_test.go to setup test storage.
promql_test.go: Copy of rules/rules_test.go.
