pkg/gate: Prefix gate metrics for selects #3154
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
kakkoyun
changed the title
simonpasquier
force-pushed
the
prefix-gate-metrics-for-selects
branch
from
722181c to
9b9d35e
Compare
kakkoyun
reviewed
Sep 11, 2020
cmd/thanos/query.go
Outdated
| engine = promql.NewEngine( | ||
| queryableCreator = query.NewQueryableCreator( | ||
| logger, | ||
| extprom.WrapRegistererWithPrefix("thanos_query_concurrent_selects_", reg), |
There was a problem hiding this comment.
Right now, this registry only used for concurrent selects. However, it could be used for others as well. I think we should wrap the registry this early.
There was a problem hiding this comment.
Good point. I'll think about it.
simonpasquier
force-pushed
the
prefix-gate-metrics-for-selects
branch
2 times, most recently
from
fff2500 to
430b3ed
Compare
Thanos query exposed `gate_queries_in_flight` and `gate_duration_seconds_bucket` metrics for concurrent selects. These metrics are now prefixed by `thanos_query_concurrent_selects_`. Signed-off-by: Simon Pasquier <[email protected]>
This change deprecates the gate.(*Keeper) struct. When Keeper is used to create several gates, the metric tracking the number of in-flight metric isn't meaningful because it is hard to say whether requests are being blocked or not. As such the `thanos_query_concurrent_selects_gate_queries_in_flight` is removed. The following metrics have been added to record the maximum number of concurrent requests per gate: * `thanos_query_gate_queries_max` * `thanos_bucket_store_series_gate_queries_max`, previously known as `thanos_bucket_store_queries_concurrent_max.` * `thanos_memcached_getmulti_gate_queries_max` Signed-off-by: Simon Pasquier <[email protected]>
simonpasquier
force-pushed
the
prefix-gate-metrics-for-selects
branch
from
430b3ed to
0a6b36a
Compare
Signed-off-by: Simon Pasquier <[email protected]>
kakkoyun
reviewed
Sep 14, 2020
| // gate.(*Keeper).NewGate only once, it is recommended to use gate.New() | ||
| // instead. Otherwise it is recommended to use the | ||
| // github.com/prometheus/prometheus/pkg/gate package directly and wrap the | ||
| // returned gate with gate.InstrumentGateDuration(). | ||
| type Keeper struct { |
There was a problem hiding this comment.
Is this still used anywhere in the codebase? If it's not the case, why not remove it directly? I don't think we have any downstream dependents.
There was a problem hiding this comment.
unfortunately we have, I should have pointed out in the PR.
https://github.com/cortexproject/cortex/blob/f1b95c68d3af4f715b11bd33c53ebaf83260f3bb/pkg/storegateway/bucket_stores.go#L72
There was a problem hiding this comment.
and given that Thanos imports Cortex too, it creates a circular dependency hence the need to keep Keeper :)
kakkoyun
reviewed
Sep 14, 2020
Signed-off-by: Simon Pasquier <[email protected]>
kakkoyun
approved these changes
Sep 14, 2020
simonpasquier
added a commit
to simonpasquier/thanos
that referenced
this pull request
Sep 15, 2020
Signed-off-by: Simon Pasquier <[email protected]>
2 tasks
kakkoyun
pushed a commit
that referenced
this pull request
Sep 15, 2020
Signed-off-by: Simon Pasquier <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Changes
The metrics on the gates for concurrent selects were exposed without a prefix (e.g.
gate_queries_in_flightandgate_duration_seconds) unlike the gate metrics for the query API. The first commit 60950ff adds thethanos_query_concurrent_selects_prefix.After more thinking, I felt that exposing the number of in-flight queries doesn't make much sense for concurrent selects because each query creates a new gate. Thus we can have (max number of concurrent selects per query x max number of concurrent queries) selects in flight and it doesn't tell whether a query is blocked on too many selects or not. I've added another commit
722181c that removes the
thanos_query_concurrent_selects_gate_queries_in_flightmetric.The following metrics have also been added to record the maximum number of concurrent requests per gate:
thanos_query_gate_queries_maxthanos_bucket_store_series_gate_queries_max, previously known asthanos_bucket_store_queries_concurrent_max.thanos_memcached_getmulti_gate_queries_maxVerification
Tested manually.