README.md in postgres-vacuum-monitor-0.4.0 vs README.md in postgres-vacuum-monitor-0.5.0
- old
+ new
@@ -6,11 +6,11 @@
# Postgres::Vacuum::Monitor
Postgres::Vacuum::Monitor provides queries that provide information about the number of dead tuples and long running queries.
This information helps to diagnose and monitor two things:
1) That the current auto vacuum settings are working and keeping up.
-2) That there are no long running queries affecting the auto vacuuming daemon.
+2) That there are no long running transactions affecting the auto vacuuming daemon.
## Installation
Add this line to your application's Gemfile:
@@ -42,18 +42,19 @@
def report_event(name, attributes)
end
end
```
-For long running queries, the event name is `LongQueries` and the attributes are:
+For long running transactions, the event name is `LongTransactions` and the attributes are:
```ruby
{
database_name: # The name of the database.
- start_time: # When the query started .
+ start_time: # When the transaction started .
running_time: # How long has it been running in seconds.
application_name: # What's the application name that is running the query.
- query: # The offending query.
+ most_recent_query: # The last query started by the transaction
+ state: # The state of the transaction - either "active" or ""
}
```
For auto vacuum the attributes are the following:
@@ -82,12 +83,12 @@
#### Dead tuples
```SQL
SELECT percentile(dead_tuples) FROM AutoVacuumLagging facet table where appName = 'my-app' SINCE 1 DAY AGO TIMESERIES
```
-#### Long running queries
+#### Long running transactions
```SQL
-SELECT application_name, query, running_time, start_time FROM LongQueries
+SELECT application_name, state, most_recent_query, running_time, start_time FROM LongTransactions
```
#### Tables that need to be vacuumed
```SQL
SELECT uniques(table) FROM AutoVacuumLagging where appName = 'my-app' since 30 minutes ago