I`m running several python (2.7) applications and constantly hit one problem: log search (from dashboard, admin console) is not reliable. it is fine when i searching for recent log entries (they are normally found ok), but after some period (one day for instance) its not possible to find same record with same search query again. just “no results”. admin console shows that i have 1 gig of logs spanning 10-12 days, so old record should be here to find, retention/log size limits is not a reason for this.
Specifically i have “cron” request that write stats to log every day (it`s enough for me) and searching for this request always gives me the last entry, not entry-per-day-of-span-period as expected.
Is it expected behaviour (i do not see clear statements about log storage behaviour in docs, for example) or there is something to tune? For example, will it help to log less per request? or may be there is advanced use of query language.
Please advise.
This is a known issue that has already been reported on googleappengine issue tracker.
As an alternative you can consider reading your application logs programmatically using the Log Service API to ingest them in BigQuery, or build your own search index.
Google App Engine Developer Relations delivered a codelab at Google I/O 2012 about App Engine logs ingestion into Big Query.
And Streak released a tool called called Mache and a chrome extension to automate this use case.