Repository navigation
[2.5.13] Memory leak during normal operation #1589
Description
/about output:
{
"version": "v2.5.13",
"uiVersion": "v2.5.13",
"commit": "036dd5e",
"deploymentId": "01KKBMQ41SRQX84E59YHVR4EV4",
"updateAvailable": false,
"latestVersion": null,
"llmActive": false,
"llmProvider": null,
"oidcActive": false,
"license": {
"name": "AGPL-3.0-only",
"version": "v1",
"plan": "OSS",
"deploymentInfo": "Managed"
},
"mode": "Standalone",
"staging": "/data/staging",
"hotTier": "Disabled",
"grpcPort": 8001,
"store": {
"type": "Local drive",
"path": "/data/local-storage"
},
"analytics": {
"clarityTag": null
}
}Description:
During our evaluation of parseable as a centralized logging solution, we have conducted two ingestion tests, sending 20GB and 80GB of log messages into the system, then performed searches over the dataset. During this time, memory usage (RSS) of the parseable process has increased:
- during the first test (20 GB ingested): 1396 MB -> 3851MB
- during the second test (80 GB ingested): 3855 MB -> 4739 MB
This suggests that there is a slow but steady memory leak.
Activity
parmesant commentedon Mar 18, 2026
Hey @lumiera
Thanks a lot for opening this issue. For how long did you keep the server running for each test?
lumiera commentedon Mar 18, 2026
The server was running constantly (there were no restarts between the tests).
Total uptime of the server is 5 days.
The tests took around 12 hours each.
parmesant commentedon Mar 18, 2026
@lumiera are you on our slack?
https://logg.ing/community
lumiera commentedon Mar 18, 2026
No.
parmesant commentedon Mar 18, 2026
As I suspected, this is not a memory leak issue but more of a default behavior
apache/datafusion#15939
We use jemalloc which might be holding on to memory based on whatever trade-offs it is designed to do for memory allocations
It would be great if you could join our slack if you want to explore this further
parmesant commentedon Mar 18, 2026
@lumiera feel free to close the issue