Some Takeways from the talks I found interesting:

 

Instrumentation and ingestion in an instant: Demo of Grafana Alloy, Grafana Beyla, and OpenTelemetry:

Takeaways:

• What is open telemetry and how it is like a USB-C connector in a world of random charging ports 😄

• Grafana Beyla : Zero-code instrumentation: Beyla automatically captures telemetry data without requiring modifications to application code or configuration. (much more about this can be found online)

• Beyla uses eBPF to inspect application executables and the OS networking layer

• Kubernetes integration: It can listen to the Kubernetes API to decorate metrics and traces with Pod and Service metadata

 

User Story : Transforming enterprise observability with Grafana Cloud: Cyara's journey to operational excellence

Takeaways:

• With improved observability, the team at Cyara, were able to run performace testing using features like k6 to test changes before they are deployed

• Platform resource utilisation dashboarding helped with infrastructure optimisation. They were able to even monitor single “problematic” nodes that would frequently alert using single resource analysis dashboards

• This eventually helped them improve costs management through financial alerting

 

AI/ML in Grafana Cloud: Saving toil, time, and money:

• SLO/SLI alert creation out of the box with grafana cloud. This also automatically also provides alerts for fast/slow burn rates on any SLO alert created.

• ML prediction based error budgeting

• Adaptive metrics : Analyzes usage patterns in dashboards, alerts, and queries and identifies unused or partially used metrics through aggregation and recommends them to be removed.

• Grafana Asserts : a contextual root cause analysis tool. Uses AI/ML to automate correlation of related issues for faster root cause analysis. It also builds an Entity Graph by analyzing metric labels from different sources, including Prometheus exporters, service meshes, and OpenTelemetry data.

 

AWS Talk : Building a generative AI assistant for Grafana Cloud: Harnessing the power of Amazon Bedrock

• Using Amazon bedrock with logs from prometheus, the chatbot is able to provide answers in the form of context based logs and patterns.

​

A user's guide to Grafana Cloud's end-to-end IRM solution

Takeaways:

• A mature system would make sure the incident management stays as close to the telemetry data as possible. It would ideally all be in 1 place to ensure smooth incident management within context.

• Grafana metric drilldowns

• The IRM toolkit allows for automated alert timeline additions from slack, auto-AI generated summaries based on accurate timelines as well as JIRA/Zoom integrations. Any dashboard links or logs are automatically attached to the incident timeline and can be used for post mortems.