For the time being though to keep it simple access is only possible to DCS-s that are not using passwords (which is the common case as no real secrets are stored there). Remember – the standard way is to exactly tell on which IP-s / hostnames your Postgres instances are available. In short the feature means that you can provide connection information to your Distributed Consensus Store (etcd, Zookeeper, Consul) that’s powering Patroni and then pgwatch2 will periodically consult this DCS and fetch the active “to be monitored” instances listing from there. The second big feature add support for our favourite Postgres HA cluster manager – Patroni, and should especially appeal to people running Postgres in a “cloudy” manner. To get quickly started there’s also a suitable “predefined config” provided and an “DB Overview” dashboard similar to Postgres / InfluxDB. Also remember that we’re talking about the “pull” model now – previously metrics were pushed to the datastore. The choice of data storage “backends” has been widened again (previous version added PostgreSQL) – it’s now possible to leverage the very popular Prometheus TSDB to store metrics gathered by pgwatch2! But a recommendation – I’d use Prometheus only when it’s already running in your organization and would strongly consider the alternatives (Postgres, InfluxDB) if planning to only monitor Postgres and starting a fresh monitoring setup as Prometheus is not able to store string values which is not optimal for Postgres monitoring. Highlights – Prometheus and Patroni support So here is a quick overview on changes – for the “full monty” and when updating from a previous version please turn to the GitHub changelog link below.
New version is incremented to 1.6 but continuing the naming tradition I’m also calling it “Feature Pack 5” as it’s mostly a feature release (with a couple of bugfixes as well). After almost half a year since last updates I’m glad to announce that another set of useful features and improvements have found their way to pgwatch2 – our Open Source PostgreSQL monitoring tool of choice.
0 kommentar(er)
