We have had some technical oppses and have since resolved it. Tweets have been delayed, however. We have increased the flood caps to send out the tweets faster.

This time, the issue has been mostly a mistake on our part. We have failed to take sufficient measures to keep the gladlyCast code base up-to-date and we succumbed issues concerning software regression.

In time, we should be pushing out all the tweets by midnight. We are observing the gladlyCast system closely for a couple of days.

Happy tweeting!

[All tweets have been sent out.]

Despite the fact that we already have multiple queueing and caching systems running for gladlyCast, the last few days we have also seen gladlyCast went down for a few minutes a few times.

Well, that has taught us one thing:
Queuing and caching systems are only good for ensuring minimum or zero lost of tweets (all your text messages are always accounted for), it, however, doesn’t help in ensuring minimum system downtime.

For that, we have just written an extensive monitoring daemon, almost similar to that of Apache HTTP Server, to ensure that gladlyCast is up most of the time.

Sorry for the pretty technical post. We promise we will not do this often. If you do not get it, fret not, it simply just means we have just greatly increased the reliability of gladlyCast.

By the way, do stay tuned. We have some announcements to be made very soon.