VictoriaMetrics/lib
Zakhar Bessarab 8198e7241d
lib/mergeset: add sparse indexdb cache (#7269)
Related issue:
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/7182

- add a separate index cache for searches which might read through large
amounts of random entries. Primary use-case for this is retention and
downsampling filters, when applying filters background merge needs to
fetch large amount of random entries which pollutes an index cache.
Using different caches allows to reduce effect on memory usage and cache
efficiency of the main cache while still having high cache hit rate. A
separate cache size is 5% of allowed memory.

- reduce size of indexdb/dataBlocks cache in order to free memory for
new sparse cache. Reduced size by 5% and moved this to a separate cache.

- add a separate metricName search which does not cache metric names -
this is needed in order to allow disabling metric name caching when
applying downsampling/retention filters. Applying filters during
background merge accesses random entries, this fills up cache and does
not provide an actual improvement due to random access nature.

Merge performance and memory usage stats before and after the change:

- before

![image](https://github.com/user-attachments/assets/485fffbb-c225-47ae-b5c5-bc8a7c57b36e)

- after

![image](https://github.com/user-attachments/assets/f4ba3440-7c1c-4ec1-bc54-4d2ab431eef5)

---------

Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
(cherry picked from commit 837d0d136d)
2024-10-24 12:43:06 -03:00
..
appmetrics all: add -metrics.exposeMetadata command-line flag, which can be used for adding TYPE and HELP metadata for metrics exposed at /metrics page 2023-12-19 03:26:02 +02:00
auth vmselect: add support of multi-tenant queries (#6346) 2024-10-01 16:37:18 +02:00
awsapi lib/awsapi: properly assume role with webIdentity token (#5495) 2023-12-20 19:07:04 +02:00
backup lib/backup/s3remote: add retryer configuration (#6747) 2024-08-07 16:59:23 +02:00
blockcache all: consistently use 'any' instead of 'interface{}' 2024-07-10 00:23:26 +02:00
bloomfilter lib: consistently use atomic.* types instead of atomic.* functions 2024-02-24 02:10:04 +02:00
bufferedwriter app/vmselect: move common http functionality from app/vmselect/searchutils to lib/httputils 2023-07-06 17:22:23 -07:00
buildinfo all: open-sourcing single-node version 2019-05-23 00:18:06 +03:00
bytesutil lib/bytesutil: smooth buffer growth rate (#6761) 2024-08-07 16:59:23 +02:00
cgroup lib/cgroup: round GOMAXPROCS to the lower integer value of cpuQuota 2024-09-23 16:11:59 +02:00
consts app/vminsert: reduce the max packet size, which vminsert can send to vmstorage 2022-04-05 15:39:58 +03:00
contextutil lib/contextutil: make golanci-lint happy by substituing unused function arg name with _ 2024-09-26 17:07:25 +02:00
decimal lib/slicesutil: add helper functions for setting slice length and extending its capacity 2024-05-12 11:33:49 +02:00
encoding lib/encoding: optimize UnmarshalVarUint64, UnmarshalVarInt64 and UnmarshalBytes a bit 2024-05-14 01:30:25 +02:00
envflag lib/envflag: do not allow unsupported form for boolean command-line flags in the form -boolFlag value 2023-08-17 13:37:05 +02:00
envtemplate allowed using dashes and dots in environment variables names (#4009) 2023-03-24 17:57:19 -07:00
envutil testing: allow disabling fsync to make tests run faster (#6871) 2024-08-30 11:18:21 +02:00
fastnum lib/fastnum: use unsafe.Slice() instead of deprecated reflect.SliceHeader 2024-02-29 17:17:24 +02:00
fasttime lib: consistently use atomic.* types instead of atomic.* functions 2024-02-24 02:10:04 +02:00
filestream vlinsert: added opentelemetry logs support 2024-09-03 20:24:01 +02:00
flagutil docs: follow-up after f0d1db81dc 2024-10-17 11:19:03 -03:00
formatutil app/vmbackupmanager: add metrics for better observability (#488) 2022-12-20 14:18:43 -08:00
fs fscore: rollback trailing space trim (#7106) 2024-09-29 14:48:36 +02:00
handshake lib/handshake: substitute time.Now() with fastttime.UnixTimestamp(), since profiling shows time.Now() is slow 2024-01-23 18:39:28 +02:00
htmlcomponents lib/htmlcomponents: use relative links for the top page and for favicon.ico 2023-11-13 20:28:17 +01:00
httpserver app/vlselect: add /select/logsql/stats_query endpoint, which is going to be used by vmalert 2024-09-06 23:00:58 +02:00
httputils app/vlinsert: support _time field without timezone information during data ingestion 2024-09-26 12:50:14 +02:00
influxutils app/{vminsert,vmagent}: add healthcheck for influx ingestion endpoints (#6749) 2024-08-05 09:45:32 +02:00
ingestserver Revert c6c5a5a186 and b2765c45d0 2024-07-03 23:57:49 +02:00
leveledbytebufferpool lib/leveledbytebufferpool: do not pool byte slices bigger than 2^18 bytes 2024-06-13 17:02:05 +02:00
logger app/vlogscli: add interactive command-line tool for querying VictoriaLogs 2024-10-01 12:24:53 +02:00
logstorage lib/mergeset: add sparse indexdb cache (#7269) 2024-10-24 12:43:06 -03:00
lrucache all: consistently use 'any' instead of 'interface{}' 2024-07-10 00:23:26 +02:00
memory all: cleanup: remove // +build ... lines, since they are no longer needed after Go1.17, and the minimum supported Go version for VictoriaMetrics source code is Go1.20 2023-11-13 19:15:42 +01:00
mergeset lib/mergeset: add sparse indexdb cache (#7269) 2024-10-24 12:43:06 -03:00
metricsql all: make fmt via the upcoming Go1.19 2022-07-11 19:23:25 +03:00
netutil lib/promscrape: fixes proxy autorization (#6783) 2024-08-19 22:50:39 +02:00
persistentqueue app/vmagent/remotewrite: follow-up for 87fd400dfc 2024-07-13 02:30:10 +02:00
procutil all: cleanup: remove // +build ... lines, since they are no longer needed after Go1.17, and the minimum supported Go version for VictoriaMetrics source code is Go1.20 2023-11-13 19:15:42 +01:00
promauth vmagent: add support of HTTP2 client for Kubernetes SD (#7114) 2024-10-08 10:37:48 +02:00
prompb Revert "Exemplar support (#5982)" 2024-07-03 16:09:18 +02:00
prompbmarshal Revert "Exemplar support (#5982)" 2024-07-03 16:09:18 +02:00
promrelabel lib/promrelabel: follow-up for 8958cecad6 2024-08-27 15:44:07 +02:00
promscrape lib/promscrape: fixed reload on max_scrape_size change (#7282) 2024-10-18 11:42:47 +02:00
promutils stream aggregation: fix possible duplicated aggregation results (#7118) 2024-09-30 14:30:34 +02:00
protoparser lib/protoparser/influx: enable batch processing by default (#7165) 2024-10-15 11:51:48 +02:00
proxy lib/promscrape: fixes proxy autorization (#6783) 2024-08-19 22:50:39 +02:00
pushmetrics lib/pushmetrics: wait until the background goroutines, which push metrics, are stopped at pushmetrics.Stop() 2024-01-16 21:18:22 +02:00
querytracer make go vet happy 2024-08-19 21:41:44 +02:00
ratelimiter app/vmagent: properly shutdown when -maxIngestionRate limit is reached 2024-04-03 02:41:11 +03:00
regexutil lib/logstorage: work-in-progress 2024-05-25 22:59:21 +02:00
slicesutil lib/slicesutil: add helper functions for setting slice length and extending its capacity 2024-05-12 11:33:49 +02:00
snapshot lib/httputils: parse URL before creating HTTP transport (#6820) 2024-08-16 11:34:49 +02:00
storage lib/mergeset: add sparse indexdb cache (#7269) 2024-10-24 12:43:06 -03:00
streamaggr tests: fix slice init length (#6897) 2024-08-30 11:18:21 +02:00
stringsutil all: consistently use stringsutil.JSONString() for formatting JSON strings with fmt.* functions instead of using "%q" formatter 2024-07-17 14:01:37 +02:00
syncwg all: open-sourcing single-node version 2019-05-23 00:18:06 +03:00
tenantmetrics vmselect: add support of multi-tenant queries (#6346) 2024-10-01 16:37:18 +02:00
timerpool lib/timerpool: use timer pool in concurrency limiters 2019-05-28 17:30:10 +03:00
timeutil vmselect: add support of multi-tenant queries (#6346) 2024-10-01 16:37:18 +02:00
uint64set lib/uint64set: optimize Set.Has() for nil Set - it should be inlined now 2024-07-16 00:00:46 +02:00
vmselectapi lib: consistently use atomic.* types instead of atomic.* functions 2024-02-24 02:10:04 +02:00
workingsetcache lib: consistently use atomic.* types instead of atomic.* functions 2024-02-24 02:10:04 +02:00
writeconcurrencylimiter app/vmagent/remotewrite: clarify the reason behind the default value for -remoteWrite.queues in the same way as the reason for -maxConcurrentInserts is defined at 73f5fb0f0c 2024-03-06 13:57:53 +02:00