On Wed, May 6, 2020, at 03:46, Stuart Clark wrote:
> The other big reason to split into multiple servers is organisational rather
> than technical.
We run Prometheus Alertmanager Grafana sets on the load balancers for each of
our environments (DEV/TEST/PROD). This allows us to, for example, re
On Wed, May 6, 2020 at 9:46 AM Stuart Clark
wrote:
> The other big reason to split into multiple servers is organisational
> rather than technical.
>
> Having servers per team or per service allows them to be owned by the same
> people as the service, rather than by a central team. For larger
> o
The other big reason to split into multiple servers is organisational rather
than technical.
Having servers per team or per service allows them to be owned by the same
people as the service, rather than by a central team. For larger organisations
this can give a lot more control over monitoring
Yeah, at some scale it becomes common to run multiple or even many
Prometheus servers, segmented by different aspects like team, service,
datacenter, sharding level, ... On the other hand, many organizations start
out with a single Prometheus server and only split things up as that one
becomes too
Hello,
did I understood correctly that due to Prometheus being very light-weight
(unlike an Elasticsearch) and efficient but having an upper-limit of xx
millions of time series per instance it is recommended to have one
Prometheus server/container per observed system (may it be an application
5 matches
Mail list logo
