Identifying Bottlenecks in Your Odoo System: Diagnostic Guide

The Misdiagnosis Problem

Your Odoo is slow. But you have no idea why.

Is it the database? Is it code? Is it the server? Your team points fingers. Your IT vendor sends a 200-page system health report you don't understand. Your CFO asks when you're getting "faster Odoo."

The problem isn't Odoo. The problem is you have no diagnostic process.

The Three Bottleneck Categories

Every slow Odoo instance has a bottleneck in exactly one of three places: the Python application layer, the PostgreSQL database, or the infrastructure (servers, RAM, CPU, disk). Yet companies almost always guess wrong.

This is why diagnosis comes first. You cannot optimize what you cannot measure. And you cannot measure what you don't look at.

The 7-Step Diagnostic Process

Step 1: Establish Your Baseline (5 Minutes)

Open a terminal. SSH into your Odoo server. Run these three commands:

top -b -n 1 | head -20
df -h
ps aux | grep odoo | wc -l

This tells you:

Write this down. This is your baseline. If nothing changes and your system is still slow, it's not infrastructure.

Step 2: Enable Query Logging (Immediate)

You are right now running database queries that take 3-8 seconds each. You don't know which ones. Fix that.

# Edit your Odoo config file (/etc/odoo/odoo.conf)
log_level = debug_sql

# Restart Odoo, wait 5 minutes, check logs
tail -f /var/log/odoo/odoo-server.log | grep -i "query time|took"

Perform a user action: create an invoice. Search inventory. Open a report. The logs will show you which SQL queries executed and how long each one took.

Look for queries exceeding 300-500ms. That is your first enemy.

DEBUG odoo.sql: SELECT res_partner.id FROM res_partner
WHERE res_partner.name ILIKE 'test' [QUERY TIME: 0.8s]

This single query takes 800ms. If someone calls it 20 times during an invoice creation, that's 16 seconds buried in your page load.

Step 3: Identify If It's Database Or Code

Open your browser. Go to your Odoo instance. Add ?debug=1 to the URL. Log in. Now open your browser's Developer Tools (press F12).

Click the "Network" tab. Create a sales order. Watch the timeline.

You'll see calls like /web/dataset/search_read or /web/dataset/call_kw. These are the expensive operations. Click one. Look at the "Time" column.

Step 4: Check For Missing Indexes

This is where 40% of slow Odoo instances live.

psql -U odoo -d my_database

-- Find unused indexes (delete them)
SELECT schemaname, tablename, indexname 
FROM pg_stat_user_indexes
WHERE idx_scan = 0 
ORDER BY idx_blks_read DESC 
LIMIT 10;

-- Find tables with excessive sequential scans (add indexes)
SELECT schemaname, tablename 
FROM pg_stat_user_tables
WHERE seq_scan > 1000 
ORDER BY seq_scan DESC 
LIMIT 10;

Common fields that need indexes:

If you find a table with 50,000+ rows being sequentially scanned thousands of times per day, adding one index will cut its query time by 60-90%.

CREATE INDEX idx_res_partner_email ON res_partner(email);

Test it. Measure again. If query time drops from 2.1s to 0.3s, you just found a $5,000+ problem you can fix in 2 minutes.

Step 5: Look For The Hidden Killer—search() Inside Loops

This is the pattern nobody sees coming.

for line in invoice_lines:
    partner = self.env['res.partner'].search([('id', '=', line.partner_id)])
    # do something

This is death. If you have 50 invoice lines, you just executed 50 database searches. If each takes 80ms, that's 4 seconds buried in code.

To find this: Enable Odoo in performance mode:

./odoo-bin --dev=performance

Watch the logs. You'll see repeated identical SQL queries. If you see the same query appearing 50+ times in a single operation, you found it. Move the search outside the loop.

Step 6: Profile Your Python Code

If your logs show fast queries but pages still load slow, Python is the problem.

odoo shell -d my_database

import cProfile
import pstats
import io

pr = cProfile.Profile()
pr.enable()

# Run your slow operation here
# e.g., self.env['sale.order'].create({'partner_id': 1, ...})

pr.disable()
s = io.StringIO()
ps = pstats.Stats(pr, stream=s).sort_stats('cumulative')
ps.print_stats(20)
print(s.getvalue())

This shows the 20 slowest function calls. If you see a custom method eating 6 seconds, that's your bottleneck.

Step 7: Check Cron Jobs (The Silent Killer)

Bad scheduled actions will strangle your entire system.

Go to Settings → Technical → Automation → Scheduled Actions.

Look for jobs set to run frequently (every 1-5 minutes). If you have 12 jobs each running every 2 minutes, they can overlap and queue up.

During your 3 PM slow period, your system might be running all at the same time:

Fix: Stagger them. Have your 2-minute jobs run at 2, 4, 6, 8... minutes. Have your 5-minute jobs run at 0, 5, 10, 15... minutes. Space them out.

The Tools That Do The Heavy Lifting

PostgreSQL's pg_stat_statements Extension

This is the single most important diagnostic tool. It logs every SQL query with execution count and timing.

# Edit postgresql.conf
shared_preload_libraries = 'pg_stat_statements'

# Restart PostgreSQL, then query it
SELECT query, calls, mean_time, max_time
FROM pg_stat_statements
WHERE database = 'my_database' 
ORDER BY max_time DESC 
LIMIT 10;

This shows your top 10 slowest queries. If you see a query taking 8.2 seconds, you have 1-2 hours of optimization work. If you see a query running 47,000 times in one hour, that's your search-in-loop problem.

Odoo's Built-In Profiler (v14+)

This is criminally underused.

The profiler shows you each SQL query executed, time taken for each, and queries grouped by operation. This gives you instant visibility. Takes 2 minutes. Catches 90% of problems.

pgBadger (For Historical Analysis)

sudo apt-get install pgbadger

# Parse PostgreSQL logs
pgbadger /var/log/postgresql/postgresql-*.log -o report.html

Open report.html in a browser. This visualizes your slowest queries, query frequency, missing indexes, and database activity timeline.

Common Diagnostic Patterns

Your Diagnostic Checklist

This Week

This Month

This Quarter

Real-World Case Study

Without diagnosis, they would have upgraded their server ($8,000+), hired more staff, or replaced Odoo.

The Final Truth About Bottlenecks

Frequently Asked Questions