Client Onboarding

Your PGFlare Session Guide

Everything you need to know before, during, and after your session — so you get maximum value on day one.

ⓘ What we need from you

You don't need to prepare extensive documentation. PGFlare reads directly from AWS via read-only IAM. The items below take most teams <30 minutes to prepare and significantly improve the depth of analysis.
01

Create the read-only IAM role (required)

PGFlare needs a read-only AWS IAM role to connect to your RDS instance. This takes about 10 minutes using our setup guide — no write permissions, no data access.

  • Follow the IAM Role Setup Guide step by step
  • Send the resulting Role ARN to [email protected] after booking
  • Permissions used: describe-db-instances, enhanced monitoring, CloudWatch read
  • Role can be revoked immediately after the engagement

02

Enable pg_stat_statements (required for Diagnostic)

This PostgreSQL extension captures query execution statistics and is essential for identifying slow queries. It's very low overhead and should already be enabled on most production instances.

  • Check: SELECT * FROM pg_extension WHERE extname = 'pg_stat_statements';
  • Enable in RDS parameter group: set shared_preload_libraries = 'pg_stat_statements'
  • Reboot required only if not already configured — allow up to 20 minutes for stats to populate

03

Note your current performance symptoms

The more context you give us before we start, the faster we can focus the analysis on what matters most to your business.

  • Which queries or endpoints are slow? (SQL text or endpoint name)
  • When did the degradation start? (date or recent AWS change)
  • Has your data volume, traffic, or schema changed recently?
  • Any recent RDS parameter group changes, major version upgrades, or snapshots?
  • Current RDS CPU %, IOPS, and connection count (from CloudWatch) if known

04

Share your instance details (helpful, not required)

We read this from AWS directly, but sending ahead speeds up the briefing call.

  • Instance class (e.g., db.r5.2xlarge)
  • PostgreSQL version
  • Storage type, size, and provisioned IOPS if applicable
  • Multi-AZ/Single-AZ
  • Any custom parameter group settings you've applied

🔑 Security reminder

Never send database credentials, passwords, or connection strings by email. The IAM role is the only access mechanism PGFlare uses. If you're unsure, reference the IAM setup guide or email us first.
Start IAM Setup →

📅 Session format

Most sessions run asynchronously — we work through the analysis independently and deliver findings by end of day. For Emergency response or when you prefer live collaboration, we run a shared Zoom/Meet session. Let us know your preference at booking.
08:00–
09:00

Briefing call (30 min, optional)

Quick call to confirm scope, priority symptoms, and any recent changes. Not required for standard Diagnostic sessions but recommended if you have specific urgency.

09:00–
11:00

Access verification & baseline data collection

We verify IAM access, confirm pg_stat_statements is populated, pull baseline metrics from CloudWatch (CPU, IOPS, connections), and run initial query analysis against pg_stat_statements and pg_stat_user_tables.

11:00–
15:00

Deep analysis

EXPLAIN ANALYZE on top queries by total_time, autovacuum config vs. table dead tuple rates, index usage analysis, lock contention review, connection pool assessment, bloat calculation. For Remediation+ sessions, remediation actions begin now in a maintenance window you specify.

15:00–
17:00

Findings report compilation

All findings ranked by business impact (cost saving + performance gain). Each finding includes severity, estimated impact, and the specific SQL or config change required. No vague recommendations — everything is actionable.

17:00

Report delivery & debrief

Full written report delivered by end of day. We're available for a 30-minute debrief call the same day or next morning to walk through findings with your team.

⚠ Emergency response differs

For Emergency (P1) engagements: we start immediately on contact, work live on the call, and target resolution in the first 2-hour block. Billed at £350/hr with a 2-hour minimum (£700).

ℹ Session fee is fixed — not time-billed

Diagnostic and Remediation+ are flat-fee sessions. The timeline above shows the typical structured working day, but you are not paying for each of those hours. You pay £497 or £897 for the deliverable — the written report and/or the applied fixes. If your instance's issues are straightforward and we deliver your report in two hours, you still pay the same flat fee. There are no time-overrun charges and no incentive to run the clock.

What your report includes

Priority Findings Ranked by Impact

Every issue is ranked by estimated monthly cost saving or P99 latency reduction so your team knows exactly what to fix first. Typical reports include 8–20 findings.

Ready-to-Run SQL and Configuration Changes

Each finding includes the specific SQL (CREATE INDEX CONCURRENTLY, VACUUM ANALYZE, ALTER SYSTEM) or parameter group value to change. Nothing to interpret — just review and apply.

Instance Right-Sizing Recommendation

Where CPU utilisation can be reduced below 50%, we recommend a specific target instance class with projected monthly saving. Includes safe-migration runbook.

Monitoring Recommendations

CloudWatch alarm thresholds for the specific metrics that matter for your workload — autovacuum lag, connection count, wait events, and query latency buckets.

Your post-session action plan

Book Your Session →
?

Do you need write access to my database?

No. PGFlare operates exclusively via read-only AWS IAM. We can read RDS metadata, CloudWatch metrics, and pg_stat_statements but cannot write to your database or read your data. For Remediation+ sessions, you apply the SQL changes — we prepare them and guide you.


?

Will any of this cause downtime?

The Diagnostic session is entirely read-only — zero risk. Remediation+ changes are designed for zero-downtime application: CREATE INDEX CONCURRENTLY rather than CREATE INDEX, parameter changes staged via RDS parameter groups rather than immediate restarts. We always flag changes that require a maintenance window.


?

What if pg_stat_statements isn't enabled?

We'll note it in the pre-session call and can still analyse table bloat, autovacuum config, index usage, and lock contention. However, query-level analysis requires pg_stat_statements and the report will be less specific on query-level findings. We strongly recommend enabling it beforehand — see the setup guide.


?

How quickly will I see improvements?

Immediate improvements are typical: adding a missing index (CONCURRENTLY) shows measurable query speedup in minutes. Autovacuum tuning typically shows CPU reduction within 24 hours. Instance right-sizing is usually possible within 2-3 weeks after applying the report findings.


?

Is there a data processing agreement (DPA)?

Yes. PGFlare operates as a GDPR-compliant data processor under UK GDPR. A signed DPA is available within 24 hours of request. See our Compliance & Security page for full details.


?

How does the flat session fee work? What if the session only takes 30 minutes?

You pay for the deliverable — not the hours. If we complete a Diagnostic in two hours because your instance's top issues are clear-cut, you still receive the same written priority report and pay £497. If a Remediation+ session runs efficiently, you still pay £897 for all agreed fixes applied and documented. There are no time-overrun charges and no minimum billable hours for flat-fee sessions. The session ends when the deliverable is complete — not when a clock runs out.


?

What if I'm not satisfied with the findings?

We stand behind the quality of our analysis. If your Diagnostic session produces fewer than 5 actionable findings, or if the projected savings don't justify the session cost, we'll extend the analysis at no charge. Email [email protected] within 48 hours of report delivery.