EC2 Network Performance & I/O Failure Diagnostics
Data strictly extracted from official AWS troubleshooting documentation. Click the symptom below to get the targeted CLI remediation commands.
🚨 Step 1: What specific error symptom are you experiencing?
Please click the most accurate description:
Quick Reference Table
| # | Scenario | Key Error Signal | Primary Fix |
|---|---|---|---|
| 1 | EBS I/O Failure / Read-Only Remount | Numerical result out of range on io_timeout | Add nvme_core.io_timeout=4294967295 to GRUB cmdline |
| 2 | conntrack Packet Loss | conntrack_allowance_exceeded via ethtool -S eth0 | Scale out instances or upgrade to larger instance type |
| 3 | DNS / IMDS / Time Sync Packet Drops | linklocal_allowance_exceeded via ethtool -S eth0 | Cache IMDS credentials; implement local DNS resolver |
| 4 | ENA SRD Fallback to Standard ENA | ena_srd_resource_utilization via ethtool -S eth0 | Scale out instances to distribute SRD connections |
| 5 | Network Interrupt Latency / C-state Stall | Deep %c6 entries in turbostat output | Add intel_idle.max_cstate=1 processor.max_cstate=1 to GRUB cmdline |