This product aims to be the metrics server in PropertyGuru. It will collect and store the metrics in a time series database, managed by Infrastructure Engineering Team. The main purpose or usage for it, for now, is mainly to act as an intemidiary to stream metrics to Datadog. Another feature of our metrics server is to do data massaging, to only send useful and clean data to Datadog.
We are deploying Prometheus on prometheus-infra
namespace to integrate with other services like istio
, kong
and collect metrics from it and forward them to datadog
as there is no direct integration between these services with datadog
and there is a need of prometheus
plugin
- Kubernetes 1.21+
- Helm 3.7+
we are installing
prometheus
using helm chart which bootstraps a Prometheus deployment on a Kubernetes cluster using the Helm package manager and with2
replicas to make sure it ishighly available (HA)
git clone https://prometheus-community.github.io/helm-charts
Along with this chart installs additional, we are deploying below dependent charts:
If we provision a new cluster and want to deploy prometheus on it then we need to checkout the kubernetes repository
- kubernetes_repo
need to create an
overlays
usingkustomize
(https://kustomize.io/) similar to the Karpenter by creating a folder with cluster name.