win5 short outages?

Was there a problem with win5 overnight? I monitored several short outages (with a 3-min monitoring interval).

No response at 1:38 AM, back at 1:41 AM
No response at 2:09 AM, back at 2:15 AM
No response at 2:24 AM, back at 2:27 AM
No response at 3:55 AM, back at 3:58 AM

(all times in EST)

Absolutely NOT

SiteUptime Stats

Please recheck your monitoring system. Alertra and SiteUptime (5min check) are the most reliable uptime checkers we’ve seen

Hyperspin is terribly inaccurate. We signed up for it only yesterday You can see the hyperspin report at:
Errors

Also, even though I consider HypeSpin very inaccurate, it has only reported 1 5min outage on Win5

[QUOTE=Yash]
Absolutely NOT
[/QUOTE]

Actually, let me be more clear (based on my logs)…

Timeout at 1:38 AM, back at 1:41 AM
Timeout at 2:09 AM, back at 2:15 AM
500 Internal Server Error at 2:24 AM, back at 2:27 AM
500 Internal Server Error at 3:55 AM, back at 3:58 AM

Tiemout is set for 5 seconds. All times are EST.

The timeout may be due to the web or the SQL server as the site is highly dependant on data from the SQL server. Do the timeout times align with backups perhaps?

As for the 500 errors, I understand that these are no uncommon but not frequent.

Our monitoring systems monitor for both timeouts and IIS errors
It will report a timeout in 30 seconds.

We haven’t recorded any problem on Win5 but we’ll keep a check on it

[QUOTE=Yash]
Absolutely NOT

SiteUptime Stats
WebsitePulse Stats

WebsitePulse actually monitors actual website content (asp/asp.net). If a site goes has any sort of asp/asp.net, db error, it rings us up

[/QUOTE]

However, WebsitePulse has a monitoring interval of 15 mins and I am using 3 minutes, so it’s not surprising that I might see something that WebsitePulse doesn’t. I am also checking actual web site content.

[QUOTE=Logan]

The timeout may be due to the web or the SQL server as the site is highly dependant on data from the SQL server. Do the timeout times align with backups perhaps?

[/QUOTE]

Yes, or if the server has seen a sudden spike in activity. But that doesn’t mean your site is down, only marginally slower for a short period of time

[QUOTE=Logan]
However, WebsitePulse has a monitoring interval of 15 mins and I am using 3 minutes, so it’s not surprising that I might see something that WebsitePulse doesn’t. I am also checking actual web site content.
[/QUOTE]

We use Nagious (stats not available live) that also checks server errors and timeouts at 2min intervals. Nothing reported

[QUOTE=Yash]
Yes, or if the server has seen a sudden spike in activity. But that doesn’t mean your site is down, only marginally slower for a short period of time
[/QUOTE]

I’ll increase my timeout to 10 seconds out of courtesy [grin], however if my site doesn’t respond in 10 seconds, I would consider that a problem.

/eddit

Timeout is the amount of time it takes for your script to execute. A 5 second period is too short IMHO. You may see very short periods where the server make take 10 to 15 seconds to execute if there is a sudden spike in activity. It does not equate to the server not responding. The server does respond with HTML

At least that is what we monitor and consider timeout.
We’ve been involved in monitoring and shared hosting for 2 years now. We know what’s acceptable and what’s achievable in a shared environment

Also, I’ve seen many monitoring software report timeouts as outages.

They can also been due to a network latency where a small network hiccup can lead to a timeout and result in a 5min outage when the server wasn’t down for even a second or had slowed down a little bit for 5 to 10 seconds