Upgrading to Zabbix 7.0.1 (or later) from Zabbix 7.0.0 with PostgreSQL/TimescaleDB results in a server crash. This issue is caused by a workaround to a compression job issue in the auditlog table in Zabbix 7.0 that irreversibly changed the compression policy of the auditlog table.
See known issues for details of fixing the auditlog table manually.
Note that there should be no such issues for upgrades from pre-7.0 Zabbix versions to 7.0.1 (or later). Follow the instructions in TimescaleDB setup page configuration section "For existing installations".
Using pg_restore
to restore a PostgreSQL/TimescaleDB backup created in Zabbix 7.0.0-7.0.4 will result in a missing base36_decode
function error, causing the restore to fail. This error occurs when restoring a backup created with pg_dump
.
See known issues for details on fixing this issue manually before creating the backup.
If you have not upgraded to Zabbix 7.0.0 yet, note that there should be no such issues for upgrades from pre-7.0 Zabbix versions to 7.0.5. For upgrade instructions, see Upgrading TimescaleDB schema.
A new index has been added to the auditlog
table to improve database and frontend response times when filtering records by Recordset ID in the Audit log.
Note that users with large audit logs may experience extended upgrade times due to the database size.
Support of TimescaleDB 1.x has been removed. The minimum required version of TimescaleDB is now 2.13.0.
A standalone Zabbix server will now stop if the database becomes read-only.
The vmware.eventlog item now also returns user information.
The net.dns.perf item now returns a response time instead of 0
when the DNS server responds with an error code (for example, NXDOMAIN
or SERVFAIL
).