#Does sentry logs offer full text search capabilities like elastisearch

9 messages · Page 1 of 1 (latest)

subtle raft
#

Does sentry logs offer full text search like elastisearch

inner forgeBOT
#

Hi @subtle raft. I'm happy to dig into this for you! I'll check my internal sources for answers to your question! :wave:

#

Short answer: not today.

  • Sentry Logs supports structured logs with text search on the log message and filtering by attributes/properties. However, requests for “full text search across all attributes/fields” are open feature requests and not currently supported like a generic Elasticsearch-style full-text index (for both Issues and Logs) (Logs docs, Make "raw search" work across all fields, Logs search to include all attributes of logs).
  • You can search raw text in the log message and query specific attributes (for example, severity, custom attributes), but broad free-text across every attribute is explicitly called out as not available; the team is gauging demand for expanding this capability (Logs docs, Logs search to include all attributes of logs).

If you need this today, consider emitting important values as structured attributes and query them directly in Sentry Logs (Logs docs).

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@subtle raft Our robots are still learning and improving, please let us know if this response was correct by reacting below.

subtle raft
#

@inner forge Then does sentry store log body in nodestore or clickhouse

inner forgeBOT
#

Let me check the knowledge base for you again. One moment! :books:

#

I’m sorry, but the knowledge sources don’t explicitly state where Sentry Logs (the new structured logs feature) store their log bodies. They do describe where traditional event payloads are stored: by default, event payloads are written to Nodestore (Postgres table nodestore_node in self-hosted), while searchable, indexed metadata lives in ClickHouse; Nodestore holds raw/unindexed blobs like contexts and breadcrumbs, and ClickHouse stores the parts you can search/aggregate (errors, spans, etc.) (Self-hosted data flow, Nodestore domain docs, Cleaning nodestore_node thread).

Given that, it’s likely logs follow a similar pattern (indexed fields in ClickHouse, raw blobs in a blob/Nodestore-like backend), but the sources provided do not confirm this for Logs. If you need a definitive answer for Logs specifically, I don’t have it in these docs. (Self-hosted data flow, Nodestore docs)

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@subtle raft Our robots are still learning and improving, please let us know if this response was correct by reacting below.

wet cosmos