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

Calculating Average Response Time

From reading through past posts in the forums, including documentation links, I've managed to piece this together:

Average response time is a rolling average of the last 5 ping attempts. Ping attempts are attempted based on the polling interval (default 120 seconds) and are stored in the database. That means that the average response time displayed for a node is actually, given the default polling interval, the average over the last 10 minutes, not the average within the last polling period. (By default -- and yes, I know you can change the display to include average and current response time, which I assume displays the rolling 10 minutes average plus the most recent average ping response)

How am I doing so far?

So, while troubleshooting a response time issue with our network team someone asked "How many ping packets are sent during the ICMP poll?" I couldn't find that data anywhere. Is it the default 5 packets, and then an average of those 5 packets?

Anyone have any ideas?

Find more posts tagged with

response_time_icmp_ping_calculation

Accepted answers

"How many ping packets are sent during ICMP poll?"

Depends:

node has status UP -> successful ICMP response = end of poll

node has status UP -> unresponsive ICMP response -> switch to fast poll (each 10 seconds ICMP request during Fast Poll Period - 2 minutes by default) = first success response ends poll or whole fast poll period.

node doesn't have status UP -> ICMP response = end of poll

Each failure ICMP request is repeated X times according your configuration (3 retries by default).

So in worst scenario when node goes from UP to DOWN status, we can send (120/10) * 3 = 36 ICMP requests. But for node which are up, or down permanently we send only 1 ICMP request.

hope this helps

All comments

"How many ping packets are sent during ICMP poll?"

Depends:

node has status UP -> successful ICMP response = end of poll

node has status UP -> unresponsive ICMP response -> switch to fast poll (each 10 seconds ICMP request during Fast Poll Period - 2 minutes by default) = first success response ends poll or whole fast poll period.

node doesn't have status UP -> ICMP response = end of poll

Each failure ICMP request is repeated X times according your configuration (3 retries by default).

So in worst scenario when node goes from UP to DOWN status, we can send (120/10) * 3 = 36 ICMP requests. But for node which are up, or down permanently we send only 1 ICMP request.

hope this helps

jbiggley

That's the detail I was looking for on ICMP response. First successful response, regardless of device state, is marked as the 'current response time' and then used to derive the average response time, based on the past 5 'current response times' recorded.

Awesome -- off to take this back to our network team to go down another rabbit hole!

jbiggley

One more question, just so I understand. (I thought I did until I started to write an email to our network team -- then realized I missed a part.)

In a node transitioning from up to down, what is the total time period Fast Polling results in if, say, the node is actually down? Does the Fast Poll period continue non-stop, or does it take a 2 minute rest during polling? (using default values of course)

ICMP Fast Poll period (2 minutes - 12 ICMP packets spaced 10 seconds apart)

+ 2 minute polling period

ICMP Fast Poll period (2 minutes - 12 ICMP packets spaced 10 seconds apart)

+ 2 minute polling period

ICMP Fast Poll period (2 minutes - 12 ICMP packets spaced 10 seconds apart)

= 10 minutes?

If the 2 minute polling periods are ignored, Fast Polling can last up to 6 minutes which could result in a node being down up to 8 minutes (node successfully polls, then node goes down immediately after, +2 minutes for next polling cycle, +6 minutes for ICMP Fast Poll)

I think this has been discussed at length elsewhere, but I thought I'd tie the two together for continuity.

Thanks

This thread will provide you all needed info I guess.

if you can't find data there, I can rephrase that once more.