Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Discard Percentage Alert

Why does NPM use error/discard count vs. error/discard percent when determining warning or critical states? Since when does the number of discards matter more than the percentage of discards?

I need to configure an alert that triggers on discard percentage over a custom timeframe. Any suggestions?

Find more posts tagged with

Accepted answers

All comments

RichardLetts

in my environment any discards are significant -- I don't really care what the percentage is (1% discards on a 100Gbps link is 1Gbps of network traffic not reaching its destination.)

Discards normally indicate some fairly significant engineering issue (e.g. bandwidth mismatch) that needs to be addressed

coen

Cisco TAC, and every other network engineer I've ever talked to about discards, always wants to know what the discard percentage is. Obviously discards represent a problem, Rich, but aren't they more, or less, of a problem when you consider actual impact? I say good luck finding a manager willing to spend thousands of dollars to upgrade the bandwidth on a link that's generating .5% discards with no user complaints. Also, if any number of discards are significant and represent fairly significant engineering issues, why isn't the critical threshold set to 1??

I have a requirement to alert on discard percentage. Any takers on best way to do that? I know reports do it and I thought about SWQL or SQL alerting, but haven't done much with that yet.

Thanks in advance for any assistance.

RichardLetts

They aren't generally used to networking at the scale we work at; In a high performance high-latency environment where we are transferring Terabytes of research data even a low level of discards will screw up the throughput of most transfer programs (few high-throughput flows). Our Genome researches often find firmware bugs in switches first, and discards (in our environment) are often due to some resource limitation (other than bandwidth), or a bug in the firmware.

Again, my environment is radically different from most corporate environments.