Thresholds and alert configurations are never enough. Period. Most alerts are setup retrospectively or using templates, and then continuously fine-tuned over time. Despite that, despite spending months and months on alerting & fine-tuning, an on-call engineer is always stressed. There’s always this fear of a black swan moment where an alert or anomaly gets missed.

​
AI Based Dynamic Alerting

DroidAgent can come in handy in such situations. It can continuously monitor your critical parts of your service, without you having to define hard boundaries on what to look for. Once you define a service in DrDroid, you can activate dynamic alerting for it from the service catalog page. By switching on AI Monitoring, the agent will continuously keep checking the data related to that service, be it logs (for an anomaly), deployments (for failures or releases), metrics (for any spike). It is designed to have back-offs and ramp-ups based on the stability (or volatility) of the data.