Troubleshooting high scale deployments

This page provides troubleshooting guidance for high scale Mattermost deployments with 100 users or more.

If these steps do not resolve the issue, and you have a paid subscription to a Mattermost offering, please reach out to our customer support team via our online support portal.

Additionally, peer-to-peer support is available for all Mattermost users in our troubleshooting forum and on our community server.

My system keeps hanging when I search for a message in Mattermost

First, check how many messages have been posted on your system, including deleted posts and posts made using automations.

Go to the System Console > Reporting > Site Statistics and review the Total Posts figure reported. If this figure is above 3,000,000 posts, we recommend deploying Elasticsearch alongside your Mattermost server for improved search performance. Follow our guides to deploy an Elasticsearch server.

How to troubleshoot server-side performance issues without Prometheus or Grafana?

Enable slow query logging in PostgreSQL and leave it enabled to gather data over time. PostgreSQL’s slow query log helps you identify queries that take longer than a specified amount of time. The slow query log isn’t enabled by default, and must be enabled manually.

To enable slow query logging globally, change the following line in postgresql.conf: log_min_duration_statement = 1000, then restart PostgreSQL. When you this value to 1000, PostgreSQL considers queries that take longer than 1 second to be slow queries and logs them in the log file. Alternatively, you can run SELECT pg_reload_conf(); to reload the configuration without a restart.

To enable slow query logging for a specific database, use ALTER DATABASE to change the configuration parameter for a single database.