Commit graph

5088 commits

Author SHA1 Message Date
Aliaksandr Valialkin
32b6ce691b
lib/storage: move common code to newRawRowsBlock() function 2022-10-21 14:46:06 +03:00
Aliaksandr Valialkin
2f8861ed9c
lib/storage: simplify code a bit after 3f5959c053 2022-10-21 14:39:44 +03:00
Aliaksandr Valialkin
1fb2be0cae
lib/{mergeset,storage}: simplify the code a bit after ae55ad8749 2022-10-21 14:33:15 +03:00
Aliaksandr Valialkin
af648279ce
lib/storage: validate timestamps in the block only if they use encoding, which needs validation
This reduces CPU usage when there is no sense in validating timestamps.

This is a follow-up for 5fa9525498

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2998
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3011
2022-10-21 00:54:37 +03:00
Aliaksandr Valialkin
edf3b7be47
lib/storage: try generating initial parts from inmemory rows with identical sizes under high ingestion rate
This should improve background merge rate under high load a bit
2022-10-20 23:27:44 +03:00
Aliaksandr Valialkin
4d71023eb9
lib/workingsetcache: increase default cache expiration from 10 minutes to 20 minutes
This increases the maximum time for cache population with new entries from 20 minutes to 40 minutes.
This

This change shouldn't increase memory usage for caches, since the prev cache cleaner
should free up memory by deleting unused prev cache as soon as possible.
See 08ca45d238 for details on prev cache cleaner.
2022-10-20 21:59:08 +03:00
Aliaksandr Valialkin
9a52b56b89
lib/workingsetcache: move the cleaner for the prev cache into a separate goroutine
This makes the code more clear after d906d8573e
2022-10-20 21:59:02 +03:00
Aliaksandr Valialkin
324e119172
lib/procutil: stop immediately after receiving the second SIGINT or SIGTERM signal
Previously VictoriaMetrics apps could stop responding to SIGINT and SIGTERM signals
if they hang for some reason in graceful shutdown procedure.
2022-10-20 21:58:49 +03:00
Aliaksandr Valialkin
6855de311c
lib/{mergeset,storage}: avoid unaligned 64-bit atomic operation panic on 32-bit platforms
The panic has been introduced in 68f3a02589

While at it, add padding to shard structs in order to avoid false sharing on mordern CPUs

This should improve scalability on systems with many CPU cores
2022-10-20 16:24:46 +03:00
Aliaksandr Valialkin
526bc8a8b0
lib/workingsetcache: drop the previous cache whenever it recieves less than 5% of requests comparing to the current cache
This means that the majority of requests are successfully served from the current cache,
so the previous cache can be reset in order to free up memory.
2022-10-20 10:48:46 +03:00
Aliaksandr Valialkin
42cda38dbc
lib/workingsetcache: use per-bucket stats counters instead of global stats counters for cache hits/misses
This should improve cache scalability on systems with many CPU cores.
2022-10-20 10:48:46 +03:00
Aliaksandr Valialkin
f22bea242f
lib/workingsetcache: randomize interval for swapping curr and prev caches
This should make CPU usage smoother over time, since different caches
will be swapped at different times.
2022-10-20 10:48:46 +03:00
Aliaksandr Valialkin
0126ea31e7
docs/CHANGELOG.md: move the BUGFIX line for 1059c4d84a into correct place 2022-10-18 20:40:37 +03:00
Nikolay
ea0596d9d8
lib/promscrape/discovery/kubernetes: correctly wrap error (#3250)
* lib/promscrape/discovery/kubernetes: correctly wrap error
follow-up after 1304824201

* Update docs/CHANGELOG.md

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2022-10-18 20:40:37 +03:00
Roman Khavronenko
f7d69c1735
vmalert: lower severity level for RW retries (#3237)
The message about dropped data still remains at `error` level.
The change supposed to make log message more clear about how
serious it is.

Signed-off-by: hagen1778 <roman@victoriametrics.com>

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2022-10-18 20:40:37 +03:00
Aliaksandr Valialkin
a408223e6d
vendor: make vendor-update 2022-10-18 10:57:06 +03:00
Aliaksandr Valialkin
574b9f4f78
app/vmselect/vmui: make vmui-update after 080562030d 2022-10-18 10:51:05 +03:00
Yury Molodov
14c7223dc4
fix: remove rounding of axis limits (#3238) 2022-10-18 10:51:03 +03:00
Aliaksandr Valialkin
d0288ea417
all: log error when environment variables referred from -promscrape.config are missing
This should prevent from using incorrect config files
2022-10-18 10:29:59 +03:00
Aliaksandr Valialkin
e4e2d1fcde
lib/protoparser/clusternative: allocate unmarshalWork after reading the data from input connection
This shortens the time when unmarshalWork is in use. This also reduces the number of unmarshalWork
objects in the pool, and its memory usage.
2022-10-18 00:24:04 +03:00
Aliaksandr Valialkin
481ca746ba
lib/protoparser/clusternative: reuse unmarshalWork in order to reduce memory allocations 2022-10-18 00:06:56 +03:00
Aliaksandr Valialkin
6f69a88a5a
lib/storage: double the number of rawRows shards on multi-core systems
This should increase data ingestion scalability on multi-core systems at the cost of slightly higher memory usage
2022-10-17 18:19:28 +03:00
Aliaksandr Valialkin
68f3a02589
lib/{storage,mergeset}: do not hold per-shard lock in fast path when adding per-shard items to the flush list 2022-10-17 18:01:55 +03:00
Aliaksandr Valialkin
c4a3d8b169
lib/promrelabel: add relabeling tests when the source label is missing 2022-10-17 14:48:29 +03:00
Aliaksandr Valialkin
00e4c7f265
vendor: make vendor-update 2022-10-14 15:20:48 +03:00
Aliaksandr Valialkin
704794f164
docs/MetricsQL.md: document that input histograms must have the same set of buckets when calculating the quantile over them
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3231
2022-10-14 14:00:26 +03:00
Aliaksandr Valialkin
e3b9815fd9
docs/CHANGELOG.md: add a note that it is recommended to use v1.82.1 instead of v1.82.0 2022-10-14 13:41:29 +03:00
Aliaksandr Valialkin
7e8c9358d1
docs/CHANGELOG.md: release v1.82.1 2022-10-14 11:34:01 +03:00
Aliaksandr Valialkin
ed324aad66
lib/bytesutil: make sure that the string passed to FastStringMather.Match() is copied before using it as a key in the internal cache map
This prevents from possible corruption of the internal cache map
when the underlying byte slice used by the string key is modified.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3227
2022-10-14 09:52:18 +03:00
Denys Holius
99ec04d6aa
k8s-monitoring-via-vm-cluster.md: remove extra spaces (#3234) 2022-10-14 09:52:16 +03:00
Yury Molodov
a50f2e141c
vmui: limit number of plotted series (#3229)
* feat: add maximum display series by tabs

* feat: add warning on PredefinedPanels.tsx

* docs/CHANGELOG.md: vmui limit number of plotted series

* docs/CHANGELOG.md: vmui limit number of plotted series

* wip

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2022-10-13 12:14:53 +03:00
Nikolay
07140e0877
lib/backup: set s3 default region to us-west-2 (#3224)
* lib/backup: set s3 default region to us-west-2
it should fix an error with region detection for bucket, if AWS_REGION env var is not set

* Update lib/backup/s3remote/s3.go

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
2022-10-13 12:06:24 +03:00
Aliaksandr Valialkin
c7360e1fa0
app/vmagent/remotewrite: typo fix after c914e4dace 2022-10-13 12:06:24 +03:00
Aliaksandr Valialkin
9bf8d64605
app/vmselect/promql: follow-up for 930f1ee153
Document the change at docs/CHANGELOG.md
Apply it to histogram_quantile() in the same way as to histogram_share()

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/pull/3225
2022-10-13 12:06:24 +03:00
Siqi Liu
c0fc716426
BUGFIX: properly calculate histogram_quantile with the same value and different string le (#3225)
Co-authored-by: 647(siki.liu) <siki.liu@huolala.cn>
2022-10-13 12:06:24 +03:00
Aliaksandr Valialkin
8f0e4901fe
docs/CHANGELOG.md: document b856581ad3 2022-10-13 10:36:13 +03:00
Dan Fredell
2152d3eb61
Multi retention standardization of docs (#3223)
* Multi retention standardization of docs

While reading through the docs the implementation details had different formatting for storageNode. This standardizes them and adds a link to the retention docs.

* Update docs/guides/guide-vmcluster-multiple-retention-setup.md

Co-authored-by: Aliaksandr Valialkin <valyala@gmail.com>
2022-10-13 10:28:39 +03:00
Aliaksandr Valialkin
207b4a6c58
app/vmagent/remotewrite: typo fix after 50f5eae0e0 2022-10-13 10:19:46 +03:00
Roman Khavronenko
895cb3e7c6
vmalert: update troubleshooting docs (#3228)
The default value of `-datasource.queryStep` has changed, so we update
the troubleshooting docs accordingly.

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2022-10-13 10:14:40 +03:00
Aliaksandr Valialkin
7a6e5f9224
lib/mergeset: mention in the error message the path to the part, which triggered the error
This should improve debuggability
2022-10-12 09:54:42 +03:00
Zakhar Bessarab
4e1c12a10d
doc: describe usage of env variables for obtaining credentials (#3219) 2022-10-12 09:33:16 +03:00
Aliaksandr Valialkin
3a14d00989
docs/CHANGELOG.md: document a4975ace86
The original commit, which led to the issue - 877940a131
2022-10-12 09:33:16 +03:00
Aliaksandr Valialkin
a0f3247b14
app/vmselect/promql: properly handle zero and negative values for -search.maxMemoryPerQuery
This is a follow-up for 04a05f161c

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3203
2022-10-12 09:33:16 +03:00
Aliaksandr Valialkin
4b7ca0597d
docs/CHANGELOG.md: document 9544b5cacf7446203fac19eb7c575779dc9b280e 2022-10-12 09:33:16 +03:00
Roman Khavronenko
7fc812d3c4
vmalert: revert unexpected fileds rename during refactoring (#3222)
Due to auto-refactoring, the filed `state` was automatically
renamed to `ruleState` when the entity with the same name
was renamed in other file. Reverting the change.

https://github.com/VictoriaMetrics/helm-charts/issues/391
Signed-off-by: hagen1778 <roman@victoriametrics.com>

Signed-off-by: hagen1778 <roman@victoriametrics.com>
2022-10-12 09:33:16 +03:00
Aliaksandr Valialkin
54e5806d54
vendor: make vendor-update 2022-10-10 22:04:09 +03:00
Aliaksandr Valialkin
d6adc977b8
docs/vmbackup.md: run make docs-sync after 076e721c22 2022-10-10 21:58:06 +03:00
Aliaksandr Valialkin
098c9bda27
app/vmselect: return back the logic for limits the amounts of memory occupied by concurrently executed queries if -search.maxMemoryPerQuery isn't set
This is needed for preserving backwards compatibility with the previous releases of VictoriaMetrics.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3203
2022-10-10 21:54:18 +03:00
Aliaksandr Valialkin
d6a65b1f44
docs/CHANGELOG.md: document e384d88abf 2022-10-10 21:54:18 +03:00
Howie
d9abdc57d4
fix issue#3053 (#3182)
vmalert: prevent duplicating label `alertname` for notifications

The issue has no impact on alerting procedure. But still needs to be fixed
for clarity. 

https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3053

Signed-off-by: lihaowei <haoweili35@gmail.com>
2022-10-10 21:54:18 +03:00