TO TAS Tile Architecture:
The Tanzu Observability TAS Nozzle consists of three main components.
- Healthwatch Exporters – HealthWatch exporters scrape metrics from TAS firehose
- Telegraf – Telegraf is an open source, light-weight server process capable of collecting, processing, and aggregating metrics
- Wavefront Proxy - A Wavefront proxy ingests metrics and forwards them to the Wavefront service in a secure, fast, and reliable manner
The Healthwatch exporters are deployed as VMs which stream metrics from the TAS Firehose. These metrics are scraped by Telegraf and converted into the Wavefront metric format. The metrics are then streamed to the Wavefront Proxy which in turn sends metrics to the Tanzu Observability cloud.
There is a separate VM installed for each Healthwatch exporter to stream metrics from the firehose along with one VM each for Telegraf and proxy.
The following is a list of VMs that get installed by default as part of the Nozzle v4 tile installation.
- TAS Counter Exporter
- TAS Gauge Exporter
- TAS Timer Exporter
- TAS SLI Exporter
- Bosh Health Exporter
- Bosh Deployments Exporter
- Cert Expiration Exporter
- Telegraf
- Wavefront Proxy
TO TAS Tile Customizations:
Change Metrics scrape Interval:
You can change the scrape interval for metrics if you want to send the metrics at higher interval. By default, it is set to 15 secs.
Scrape interval can be changed in Telegraf Agent Config tab as shown below.
Sending TAS data via Proxy deployed outside of Tile:
There is an option in tile that allows you to send data via external proxy. This is helpful for customers who have central monitoring/observability teams that require all data to be sent via specific production proxy(s).
Follow the steps below to send data via external proxy:
- Click Telegraf Agent Config
- Select “Yes” in Advanced Options
- Provide proxy URL or IP in the Custom Proxy URL (This could also be a load balancer URL if external proxies are behind load balancer)
- Provide proxy port in Proxy Port (Default is 2878)
Customize Platform Metrics Ingestion into TO:
You can choose not to ingest certain platform metrics into TO if you do not care about monitoring them. For example, if you don’t want to ingest certificate expiration metrics into TO then you can assign no VM instance to Cert Expiration Exporter and all the metrics that this exporter scrapes will not get ingested.
- Click Resource Config
- Set zero instances for the exporter that you don’t want metrics to be emitted
Customize Proxy Configuration Settings:
You can customize proxy configuration settings and create add pre-processor rules to alter/filter metrics on proxy.
- Click Wavefront Proxy Config
- Click Custom under Custom Proxy Configuration
- Add custom proxy configurations in Config box and pre-processor rules in Pre-processor Rules box
Comments