[rabbitmq-discuss] Rabbitmq help

[Michael Klishin]

On sep 17, 2013, at 5:00 a.m., sagu prf <sagu.prf1 at gmail.com> wrote:

> how this shovel act upon one DC failure /Network failure ?

Shovel is AP in CAP, it will try to reconnect and republish messages.

> 
> How i can bounce the service gracefully in this scenario ?

If you need replication with more than one point, consider federation or
perhaps a load balancer in front of Shovel.

> How to handle the issue  if any the Rabbitmq are overloaded ?

Below is a very generic advice without knowing any details about
your workload.

Add more nodes to the cluster (Shovel can replicate between clusters),
make some clients connect to the new node and use additional queues there.
More nodes will not help you if the issue is a single giant queue (queues are
not sharded: this may be implemented in a future version).

You can monitor queue lengths and message rates using HTTP API and
add more nodes to the cluster when you consider it overloaded.

MK



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 495 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130917/4d45d2db/attachment.pgp>