The instrumentation should be disabled during oddball runs such as rake tasks or deploys and such. Should also abide by the NewRelic global ENV['NEWRELIC_ENABLE'].

This may be specific to running fakeredis, but I was seeing

ArgumentError: negative argument
from .../newrelic-redis-2.0.0/lib/newrelic_redis/instrumentation.rb:75:in `*'

This was happening when doing a multi/exec block, because the args list on those commands is empty.

It would be very cool to be able to ignore specific methods (like BRPOP). BRPOP is going to be very high response time and throughput on sidekiq workers, by nature, and can shadow out other database items in % utilization.

If I end up building this, I'll be sure to PR it.

I just released redis-rb 4.0.0 RC1.

Can you please test against it and report back?

Thank you!

Hi

I just installed the gem newrelic-redis on a project using redis, resque and resque-scheduler. Resque workers are starting fine but the resque-scheduler is getting an undefined method error (the trace log can be found at the end of this issue)

I changed my resque.rake such a in this issue (http://stackoverflow.com/questions/9961044/postgres-error-on-heroku-with-resque/9964924#9964924) but I am still getting the error

2012-10-31 12:17:54 Loading Schedule
2012-10-31 12:17:54 Scheduling move_day_pinned_tasks 
2012-10-31 12:17:54 Scheduling move_week_pinned_tasks 
2012-10-31 12:17:54 Scheduling move_month_pinned_tasks 
2012-10-31 12:17:54 Scheduling send_task_reminders_30_mins_interval 
** Invoke resque:scheduler (first_time)
** Invoke resque:scheduler_setup (first_time)
** Execute resque:scheduler_setup
** Execute resque:scheduler
rake aborted!
undefined method `trace_execution_scoped' for Redis::Client:Class
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/newrelic-redis-1.3.2/lib/newrelic_redis/instrumentation.rb:33:in `call_with_newrelic_trace'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/redis-2.2.2/lib/redis.rb:681:in `block in del'
/usr/local/rvm/rubies/ruby-1.9.3-p125-perf/lib/ruby/1.9.1/monitor.rb:211:in `mon_synchronize'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/redis-2.2.2/lib/redis.rb:680:in `del'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/redis-namespace-1.0.3/lib/redis/namespace.rb:213:in `method_missing'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/resque-scheduler-2.0.0.g/lib/resque/scheduler.rb:103:in `load_schedule!'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/resque-scheduler-2.0.0.g/lib/resque/scheduler.rb:45:in `run'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/resque-scheduler-2.0.0.g/lib/resque_scheduler/tasks.rb:23:in `block (2 levels) in <top (required)>'
/var/www/my_project_staging/shared/bundle/ruby/1.9.1/gems/rake-0.9.2.2/lib/rake/task.rb:205:in `call'

As per instructions I added this gem to my Gemfile and run bundle install. No it should start tracking redis but unble to see any log on new-relic.

Could someone please help me ? and let me know what is the correct way to make it work.

Also please let me know where I can see redis logs on NewRelic dashboard.

Looking at the core rpm instrumentation for mongo, I see it is added under the "Datastore" namespace:
https://github.com/newrelic/rpm/blob/master/lib/new_relic/agent/datastores/mongo/metric_translator.rb#L54

While redis is added under "database":

https://github.com/evanphx/newrelic-redis/blob/master/lib/newrelic_redis/instrumentation.rb#L67

I wonder if simply changing the namespace is all that is needed, or there is more to it?

Basically what I want is to have the redis issues pile up in the "Slow SQL Traces" block of the UI.

Before we migrated from memcache to redis NewRelic was showing it as separated color on App Server Response Time graph on Dashboard.

But redis is not shown there. I believe it is included as part of ruby or database.

Is it possible to make distinguish it from the rest just like memcache?

Quoting https://docs.newrelic.com/docs/apm/apis/application-examples-v2/obtaining-web-transaction-response-time-data-v2#datastore-v-database

As of February 2015, the New Relic agents (starting with Python) will begin to display the database time broken out into specific database types. Agents will replace the Database metric name with Datastore/${DBTYPE}, where ${DBTYPE} will identify the appropriate database; for example, Postgres, MongoDB, MSSQL, ODBC, Redis, SQLite, Memcached, etc.

Would it be possible to make use of Datastore/${DBTYPE} in the newrelic-redis gem? It sounds like it's an all around improvement over Database/${DBTYPE}.

Would it be possible to silence the "Installing Redis instrumentation" logs when New Relic is disabled? i.e rake tasks, development mode etc.

Was doing some research on third-party extensions to the Ruby agent (newrelic_rpm), and I happened across #18.

There's actually work needed on the New Relic UI side to enable the Datastore functionality. These changes alone won't work until a new (as yet unreleased) Ruby agent is available.

Also, it looks like the metric names will need to be tweaked to line up with what the New Relic UI is expecting.

Given that, could we hold up on releasing newrelic-redis with these changes until the new Ruby agent is out? I'll get back to you when that's available, along with the finalized metrics to record then.

I don't get any New Relic profiling for Redis in my Sinatra app. Should it work? If so what am I missing?

I have the following in my Gemfile:

gem 'newrelic_rpm', '3.4.2.1'
gem 'newrelic-redis', '1.3.2'

And the following in app.rb:

configure :production do
  require 'newrelic_rpm'
end

Redis-rb requires &block to support the call args. needs passing on lines 23 & 37 of instrumentation.rb