Scenerio Based Question
Q:How would you troubleshoot performance issues in a ServiceNow instance?
Answer: If slow performance is experienced by all users and affects all applications in the instance, check these common symptoms, causes and remediation steps.
Test network connectivity 📶 — Verify there are no WAN/ISP/firewall issues between users and the ServiceNow instance.
Tip: run
Check JVM memory 🧠 — Ensure the application server JVM has free heap and isn't suffering from frequent GC pauses.
ping, traceroute, or check your corporate network logs and CDN status. Issues here cause perceived slowness even if the instance is healthy.Tip: monitor heap usage, GC frequency/pauses, and growth trends. Abnormal memory usage can block transactions and slow response times.
Inspect disk I/O on the DB server 💽 — High read/write latency or saturation on database storage will slow queries and page loads.
Tip: check IOPS, queue length and latency metrics. If storage is the bottleneck, consider tuning queries, adding indexes, or working with DB infra to improve IO.
Verify CPU resources 🔥 — Confirm application and DB servers have adequate CPU to meet demand.
Tip: sustained high CPU can indicate runaway jobs, inefficient scripts, or high concurrency. Identify the offending processes and optimize them.
Check transaction concurrency 🔁 — High transaction concurrency can overload DB locks and threads, causing timeouts.
Tip: look for spikes in active transactions, long-running queries, and blocked sessions. Throttle or run heavy jobs asynchronously if needed.
🔎 Additional investigation steps
- Review system logs and ServiceNow Performance Dashboard for slow transactions and hotspots.
- Look for recent deployments, integrations, or scheduled jobs that correlate with the timing of slowness.
- Use GlideRecord / slow query analysis to find expensive queries and add indexes where appropriate.
- Consider enabling debug/profiler for targeted components, and reproduce with a test user to capture traces.
When to escalate
If infrastructure metrics (network, disk, CPU) are saturated or DB vendor involvement is required — escalate to platform/infrastructure teams or ServiceNow support.
Quick wins
- Identify & kill runaway scheduled jobs
- Optimize slow Business Rules, Script Includes, and WS calls
- Cache or paginate heavy reports
💡 Short answer to the interviewer: Run a full-stack health check — network → JVM → disk I/O → CPU → concurrency — and then drill down into slow transactions and heavy jobs. Fixing the root cause (queries, scripts, infra bottlenecks) is the goal, not just mitigating symptoms.