# DrDroid: How AI SRE Helps Engineers who are on-call for production monitoring

## Day 0 — Incident Response & Firefighting

*Something is broken right now. You need answers fast.*

| **#** | **Your Task** | **How Doctor Droid Helps You** |
| --- | --- | --- |
| 1 | **"Our API latency spiked, what changed?"** | Ask the agent — it pulls recent deployments from ArgoCD/GitHub Actions, checks Grafana/Datadog metrics, and shows you what changed around the time latency spiked. No need to open 5 tabs. |
| 2 | **"Which pods are crashing in production?"** | Ask the agent — it lists failing pods, pulls their logs, shows recent K8s events, and surfaces restart counts. You get a full picture in one response. |
| 3 | **"Is the database the bottleneck?"** | Ask the agent — it runs slow query analysis on your Postgres/MySQL, checks connection pool usage, and correlates with application error rates from Datadog or New Relic. |
| 4 | **"I got paged, what's actually going on?"** | The agent auto-investigates when an alert fires. By the time you open it, there's already a summary with metrics, logs, and likely root cause pulled from your connected sources. |
| 5 | **"Are other services affected too?"** | Ask the agent — it checks health across your connected sources (Grafana, CloudWatch, K8s, Datadog) and tells you which services are degraded vs healthy. |
| 6 | **"I need to check CloudWatch logs for this error"** | Tell the agent the error pattern and time range — it queries CloudWatch Logs, Loki, or Elasticsearch directly and returns matching entries. No console login needed. |
| 7 | **"Run this PromQL/NRQL query for me"** | Give the agent your query — it executes against Prometheus, Grafana, New Relic, or Datadog and returns results inline. Great for quick checks during incidents. |
| 8 | **"SSH into the box and check disk space"** | Tell the agent the command — it runs Bash commands on remote hosts via SSH and returns output. No need to find the SSH key or remember the hostname. Works even when you don’t have laptop access. |
| 9 | **"Notify the team on Slack about this outage"** | Tell the agent what to post and where — it sends a formatted message to your Slack channel with the context you provide. |
| 10 | **"Escalate this to PagerDuty"** | The agent creates or escalates a PagerDuty/OpsGenie incident based on the investigation findings. You don't need to context-switch to the PagerDuty UI. |
| 11 | **"Create a JIRA ticket for the post-mortem"** | Tell the agent the summary — it creates a JIRA ticket with the incident details, investigation findings, and relevant links. |
| 12 | **"Did the last deploy cause this?"** | Ask the agent — it checks the latest GitHub PR merges, Jenkins builds, ArgoCD sync status, and correlates timestamps with when the issue started. |
| 13 | **"Roll back the deployment"** | Tell the agent to trigger a rollback pipeline — it kicks off the Jenkins build or GitHub Actions workflow you specify. |
| 14 | **"Check if this API endpoint is responding"** | Give the agent the URL — it makes an HTTP call and tells you the status code, response time, and body. Works for any internal API. |

## Day 1 — Operational Tasks & Maintenance

*No fire, but you need to keep things running smoothly.*

  

| **#** | **Your Task** | **How Doctor Droid Helps You** |
| --- | --- | --- |
| 1 | **"What's the current state of our K8s cluster?"** | Ask the agent — it shows pod status across namespaces, node resource usage, recent events, and any pods in CrashLoopBackOff or Pending state. |
| 2 | **"Are all our data sources healthy?"** | The agent tests connectivity to every configured connector every 10 seconds. Ask it for the current status, or set up Slack alerts for failures. |
| 3 | **"Show me all Grafana dashboards we have"** | Ask the agent — it returns the full inventory of dashboards, datasources, and folders it auto-discovered. Same for Datadog monitors, K8s resources, DB schemas, etc. |
| 4 | **"How big are our database tables getting?"** | Ask the agent to run a table size query on your Postgres/MySQL/ClickHouse — it returns sizes, row counts, and index usage. |
| 5 | **"Any slow queries running right now?"** | Ask the agent — it checks pg\_stat\_activity or equivalent on your database and shows long-running queries with their duration and state. |
| 6 | **"Check if our Jenkins pipelines are green"** | Ask the agent — it pulls recent build status from Jenkins and tells you which jobs passed, failed, or are stuck. |
| 7 | **"Is our ArgoCD app in sync?"** | Ask the agent — it checks sync status and health for your ArgoCD applications. Flags any that are out-of-sync or degraded. |
| 8 | **"Pull logs from the payment service for the last hour"** | Tell the agent the service and time range — it queries Loki, CloudWatch, or Elasticsearch and returns the logs. No need to remember log group names. |
| 9 | **"What GitHub PRs were merged today?"** | Ask the agent — it queries your GitHub repos and lists merged PRs with authors, titles, and timestamps. |
| 10 | **"Trigger a build for the staging environment"** | Tell the agent which Jenkins job or GitHub Actions workflow to run — it triggers it and reports back the status. |
| 11 | **"Send a daily cluster health report to Slack"** | Set up a scheduled playbook — the agent runs K8s health checks daily and posts a summary to your Slack channel automatically. |
| 12 | **"Check MongoDB replica set health"** | Ask the agent — it runs the appropriate commands against your MongoDB instance and returns replica set status and lag. |
| 13 | **"List all CloudWatch alarms in ALARM state"** | Ask the agent — it queries CloudWatch and returns currently firing alarms with their metric details and thresholds. |
| 14 | **"What Datadog monitors are in alert?"** | Ask the agent — it checks your Datadog monitors and lists any that are currently alerting or warning. |
| 15 | **"Run this custom SQL query on production"** | Give the agent the query — it executes against your connected Postgres, MySQL, ClickHouse, or BigQuery and returns results as a table. |
| 16 | **"Check disk and memory on our VMs"** | Tell the agent to run df -h and free -m via Bash — it SSHs in and returns the output. |
| 17 | **"Update this JIRA ticket with today's progress"** | Tell the agent the ticket ID and comment — it adds the update to JIRA without you opening the browser. |
| 18 | **"What new resources were created in our K8s cluster this week?"** | Ask the agent — it compares the current asset inventory with the previous discovery and highlights new pods, services, or deployments. |

  

* * *

## Day 2 — Automation, Optimization & Reliability

*You want to stop doing repetitive things and build resilience.*

  

| **#** | **Your Task** | **How Doctor Droid Helps You** |
| --- | --- | --- |
| 1 | **"Automate our incident response runbook"** | Build a runbook as per your internal process: when alert fires → pull metrics from Grafana → check K8s pods → query DB → post findings to Slack. Runs automatically on every alert. |
| 2 | **"Auto-restart pods when OOM detected"** | Set up a workflow: if K8s event shows OOMKilled → restart the pod → notify on Slack → log to JIRA. No human in the loop. |
| 3 | **"Scale up when CPU crosses 80%"** | Create a conditional workflow: check CPU metrics from Prometheus/CloudWatch → if above threshold → trigger scaling via K8s or API call → confirm on Slack. |
| 4 | **"Track our SLOs across services"** | Set up scheduled playbooks that query Prometheus or Datadog for error rate and latency → calculate against SLO targets → post weekly reports to Slack. |
| 5 | **"Alert me before we hit our error budget"** | Build a workflow that checks error budget consumption daily → if > 80% consumed → post warning to Slack and create a JIRA ticket. |
| 6 | **"Correlate deploys with performance changes"** | Set up a playbook that triggers after every deploy → compares pre/post metrics from Grafana/Datadog → flags regressions automatically. |
| 7 | **"Stop SSHing into boxes for the same checks"** | Convert your common SSH commands into playbooks — disk space, process checks, log tailing. Run them from Doctor Droid with one click or on schedule. |
| 8 | **"I keep opening 5 dashboards for the same investigation"** | Build a single playbook that queries all 5 sources and gives you a combined view. Next time, ask the agent instead of opening dashboards. |
| 9 | **"Validate infrastructure after every Terraform apply"** | Set up a post-deploy playbook: check K8s resources → verify CloudWatch alarms exist → test endpoints → report pass/fail. |
| 10 | **"Audit our K8s RBAC and network policies weekly"** | Schedule a playbook that lists RBAC bindings and network policies → compares against expected state → flags drift on Slack. |
| 11 | **"Auto-create a JIRA ticket when a deploy fails"** | Build a workflow: monitor Jenkins/GitHub Actions → if build fails → create JIRA ticket with build logs and assign to the team. |
| 12 | **"Map out which services talk to which"** | Enable Network Mapper — it discovers service-to-service communication in K8s and shows you the dependency graph. |
| 13 | **"Check if our Datadog monitors match our runbook"** | Ask the agent to list all Datadog monitors — compare with your documented expectations. Set this up as a weekly drift check. |
| 14 | **"Reduce alert fatigue for the team"** | Use alert grouping and conditional workflows — similar alerts get grouped, only actionable ones reach Slack/PagerDuty. |
| 15 | **"Automate capacity reporting for leadership"** | Schedule a monthly playbook: query CloudWatch/K8s for resource trends → query BigQuery for cost data → format and send via email or Slack. |
| 16 | **"Clear application cache when memory exceeds threshold"** | Build a workflow: check memory metrics → if above limit → make API call to flush cache → verify memory dropped → notify on Slack. |
| 17 | **"Run synthetic health checks every 5 minutes"** | Schedule a playbook that hits your critical endpoints via HTTP, checks response codes and latency, and alerts on Slack if anything degrades. |
| 18 | **"Onboard new services faster"** | When a new service deploys, the agent auto-discovers it in K8s, finds its Grafana dashboards and Datadog monitors, and catalogs everything. You see it in your inventory immediately. |

## **What's Next?**

DrDroid helps your team move from **reactive firefighting** to **proactive operations**. Whether you're debugging an incident at 3 AM or building automation to prevent the next one, DrDroid becomes your team's operational co-pilot.

**Ready to see how DrDroid works with your stack?**

*   [Watch platform demo videos](https://www.youtube.com/@DrDroidDev)
    
*   [Check our integrations](https://drdroid.io/integrations)
    
*   [Read the documentation](https://docs.drdroid.io/)
    
*   [See customer case studies](https://drdroid.io/case-studies)
    

**Want to try it out?** Setup takes 1-2 hours and you'll see value from your first investigation. [Get started here](https://drdroid.io/).