|Subject||Re: [firebird-support] Firebird for a web application|
On Wed, Nov 18, 2015 at 3:16 PM, Alexey Kovyazin ak@... [firebird-support] <email@example.com> wrote:
What kind of monitoring are you doing with Nagios?
Do you mean this plugin? http://exchange.nagios.org/directory/Plugins/Databases/check_firebird-2Epy/details
Normally, Firebird monitoring should include:
1) Transactions monitoring, including long-running and stuck transactions
2) # of simultaneous connections, statements, transactions, and associated counters (memory, IO)
3) Temp files size and quantity
4) Lock table parameters monitoring
5) Firebird.log monitoring
6) Indices health monitoring: activated/non-activated, non-restored
7) Database files size monitoring, including delta files in case of nbackup lock
and some others.
We have written multiple java EE web applications with Firebird on linux as the database. Usually we will use Glassfish (or Payara) as the app server and run Firebird in classic or superclassic form. We will usually use Nagios to monitor the servers and databases.
My only complaint is that sometimes we can get the lock manager bogged down when we are really pounding a database with a large amount of small transactions. Otherwise we are overall pretty happy.
On Wed, Nov 18, 2015 at 2:42 PM, mirco.malaguti@... [firebird-support] <firstname.lastname@example.org> wrote:
Hi all here,
I want to ask some question about using Firebird in a production environment for a web application.
My experience is more then 12 years with Firebird for our client server application and have to say is a very good experience.
Now that application is being ported to web by a complete rewrite.
From an architectural point the database is single instance and multi-tenant, using logical tenant ids.
The application is a rather standard JavaEE one.
This means that all of our customers will have all their data together and obviously my main concerns are reliability, availability, performance and maintainability.
The application is mostly interactive in an OLTP fashion, with little or no datawarehousing.
We expect to have a production database of more or less 200 GB in size.
The system will be Linux-based hosted somewhere by our provider.
We now have to choose between Firebird and other engines, so I have to asses:
- tools available for effective remote monitoring, analysis, tuning and troubleshooting
I'm told that here standard names are nagios, cacti, SNMP and others
- experience by someone that had worked or is working on system of this kind
My intention is to stay with Firebird because my experience is positive and like to stay in an environment that take simplicity as a value, but I want to do what will need to be done well, anyway.
Any help is greatly appreciated.