Q: Your monitoring dashboard shows that a server's memory usage has been steadily climbing 1% per day for two weeks. The application developers say "it's fine." How do you approach this as an SRE?
A steady, linear increase in memory usage is a classic memory leak pattern. Even if the application functions today, at 1%/day, the server will OOM in approximately 2–3 weeks.
Want to master this scenario in a live sandbox? The Linux Foundation's Certified System Administrator (LFCS) & SRE Track covers this exact problem with hands-on terminal drills.
🛠️ Production Runbook & Step-by-Step Resolution
Direct Answer: "Never reboot a server blindly; always capture top process telemetry, lsof descriptors, and thread dumps first. The interviewer is testing: Capacity planning, trend analysis, proactive SRE.. I structure my answer around systematic triage first, root cause analysis second, and permanent remediation third."
Initial Diagnostics & Root Cause Analysis
A steady, linear increase in memory usage is a classic memory leak pattern. Even if the application functions today, at 1%/day, the server will OOM in approximately 2–3 weeks.
- Quantify the risk:
- Current usage: 70% → OOM in ~30 days
- Calculate the exact date the server will hit critical threshold (e.g., 95%)
- Present a clear timeline to developers: "This server will crash on [date] without intervention"
- Gather evidence:
- Identify the leaking process: Compare RSS (Resident Set Size) growth across processes. The one growing linearly is the culprit.
Remediation & Permanent Safeguards
SRE approach: Never dismiss a linear trend—it's a countdown timer to an outage.
# Track per-process memory growth
ps aux --sort=-%mem | head -20
# Check for growing heap
pmap -x <PID> | tail -1
# Monitor over time
pidstat -r -p <PID> 60
- Short-term mitigation:
- Schedule periodic application restarts during off-peak hours
- Set up memory-based alerts at 85% and 95%
- Implement cgroup memory limits to prevent one process from OOMing the entire server
- Long-term fix: Work with developers to profile the application using language-specific tools (Valgrind for C/C++, heap dumps for Java, memory profilers for Python/Go).
- Quantify the risk:
- Current usage: 70% → OOM in ~30 days
- Calculate the exact date the server will hit critical threshold (e.g., 95%)
Senior hiring managers and bar raisers evaluate how you prioritize during outages. Avoid these common anti-patterns:
- Issuing immediate kill -9 (SIGKILL) commands rather than allowing graceful termination via kill -15 (SIGTERM) when configuring monitoring dashboard shows SIGKILL immediately halts execution without allowing processes to flush memory buffers, commit database transactions, or remove stale PID locks.
-
Overlooking file system inode exhaustion (df -i) when diagnosing 'No space left on device' errors when configuring monitoring dashboard shows
Millions of small files or unrotated session tokens can exhaust available file system inodes even when standard disk storage (
df -h) shows ample free space. - Exhausting open file descriptor allocations (ulimit -n) under intensive concurrent socket connections when configuring monitoring dashboard shows Reaching file descriptor exhaustion throws 'Too many open files' errors, preventing services from accepting new TCP client socket connections.
-
Permitting defunct zombie processes to accumulate without fixing parent process SIGCHLD waitpid reaping when configuring monitoring dashboard shows
Zombie processes consume process table entries; accumulating zombies can exhaust the system maximum PID limit (
pid_max).