-
Notifications
You must be signed in to change notification settings - Fork 3.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add option max-tags-per-database to limit high cardinality data #7146
Comments
Would add to this that we should have alerts to the log when we start to Helps us stay proactive in cleanup and maintenance. On Thursday, August 11, 2016, Alex Davies notifications@github.com wrote:
Sebastian BorzaPGP: EDC2 BF61 4B91 14F2 AAB4 06C9 3744 7F3F E411 0D3E |
Of course if we do #7151 well (which looks great, thanks!), this becomes a non-issue; this is more of a sticking plaster until that can be done! |
A similar proposal is in #6679 |
This is important because if you accidentally load in a huge amount of high cardinality data, you can easily get into a place that InfluxDB will OOM if you attempt to delete the data,I am curious, why would InfluxDB OOM when I attempt to delete the data? |
The documentation suggests to add a unique tag for each point to avoid points from overwritten when timestamps are the same[1]. I tried this solution, and influxdb emits a warning after 100,000 insertions: WARN: 100% of max-values-per-tag limit exceeded: (100000/100000), db=testdb measurement=test tag=uniq service=store The limit can be removed[2]. But it feels like a non-sustainable hack. From my reading of the documentation[3] I understand that tags are not meant to be used in this way. The issue that introduced the limit[4] suggests that high cardinality data is bad. So, maybe I ought to try another solution... [1]: https://docs.influxdata.com/influxdb/v1.3/troubleshooting/frequently-asked-questions/#how-does-influxdb-handle-duplicate-points [2]: https://stackoverflow.com/questions/43770354/max-values-per-tag-limit-exceeded-influxdb [3]: https://docs.influxdata.com/influxdb/v1.3/concepts/schema_and_data_layout/#encode-meta-data-in-tags [4]: influxdata/influxdb#7146
Much as max-series-per-database in #7095, add a max-tags-per-database to limit high cardinality data.
Writes beyond that should be dropped and logged with a rate limit on the log (one summary log per minute or something saying x writes dropped).
This is important because if you accidentally load in a huge amount of high cardinality data, you can easily get into a place that InfluxDB will OOM if you attempt to delete the data, so your only choice is to try to move the files on disk out the way which deletes other data.
The text was updated successfully, but these errors were encountered: