From: Clara Armengou <[log in to unmask]>
Date: Tue, 24 Jul 2018 15:00:58 +0200

Dear community,

Since Wednesday 18th we have been experiencing significant slow-down on the
DOAJ in all its interfaces.  This is an outline of what we have been doing
to combat this, and the more aggressive measures that we are now going to
take to resolve the issue.

# Up to now

We have been monitoring the usual diagnostics for the datastore, which look
at memory, CPU usage, and network performance.  All of these are
registering within normal working parameters throughout the issue.

Nonetheless, on average requests are taking 3 - 5 seconds longer than they
should, and with no clear cause.  There have been brief outages too, though
for the most part the system has remained "up" in the technical sense. This
is an unprecedented event.


# Plan of action

Turn off the whole system, and re-introduce capabilities bit-by-bit until
we identify the area(s) that are responsible for the performance issues.

During this period large areas of the system will be unavailable, and this
may result in a degraded experience in particular for end users of the API
and OAI-PMH endpoint.

We hope to resolve these issues in the next day or so. We apologise for the
inconvenience and thank you for your patience.


-- 
Clara Armengou
Project and Communications Manager

*Directory of Open Access Journals*
*www.doaj.org <http://www.doaj.org> *