Site directory
Petrichor home
Environments
Six environments, one team of engineers
Development, QA, performance, sales demo, customer version and virtual server environments, built on our own hardware and operated by Petrichor engineers.
Shared development environments
Shared development services and databases, built from a configuration your team approves and maintained by Petrichor engineers on our own hardware.
QA and regression test environments
Repeatable QA stacks rebuilt from your configuration before each test cycle, operated by Petrichor engineers on our own hardware.
Performance and load-test environments
Load-test environments sized for the test you need, kept apart from production, and taken down when the results are in.
Demo environments for software sales
Stable demo environments with realistic data, reset between calls and kept apart from customer data.
Environments to reproduce customer bugs
Stand up the exact version a customer runs to reproduce and fix their issue, then take it down.
Virtual server hosting
Windows and Linux virtual servers in our Austin data center, with managed backups, replication and a move we handle, run by Petrichor engineers.
Audiences
Hosting for apps built with AI tools
You built the app with AI tools. Petrichor engineers host it on our own hardware and handle the patching, backups and monitoring around it.
QA, load test and demo environments for software teams
Clean QA stacks, load tests on a copy of production and demo environments that reset, built and run by Petrichor engineers for CTOs, developers and product managers.
Private cloud for business owners
A predictable monthly cloud bill you can read, support for HIPAA- and PCI-regulated workloads, and engineers who answer when something needs to change.
Service
How Petrichor works
Talk to a consultant, agree the spec, and Petrichor engineers build the environment on our own hardware. You choose how hands on you stay, and billing is predictable.
Security and isolation
Each Petrichor customer has its own network segment, on hardware we own in the U.S., operated by engineers you can reach.
Guides
Cloud guides
Guides on QA environments, restores, operating costs, migrations and release ownership.
Trial environments for enterprise prospects, kept away from production
Plan an enterprise software evaluation with agreed workflows, separate access and integrations, controlled changes, and a recorded closeout.
Parallel QA environments for two releases in the same week
What two QA stacks have to keep apart when a hotfix and the next release test in the same week: the database, sandbox accounts, sign in callbacks and scheduled jobs.
What enterprise buyers ask to see about your regression testing
The five items in a regression record for one release: build, environment, dataset, results and approval, with the public framework that asks for them.
How long to run a soak test
Set a soak test's length from the slowest scheduled cycle in the system, hold it at average load, and read memory, connections, disk and latency as slopes.
One demo environment per rep or one per vertical?
Count overlapping demos on the calendar, check whether one tenant can be reset alone, and multiply stacks by releases before choosing a demo environment pattern.
Where the hours go in a slow regression suite
A ledger for one regression cycle: hand rebuilds, dataset loads, fixture setup, triage and reruns, with the commands that show which hours the tests own.
Performance budgets per release
Set a release's latency budget from the run to run spread, measure the last release and the candidate on one environment, and count queries per request.
When to move QA off AWS ahead of production
Move QA off AWS first when it rebuilds often from data staged once and none of its tests call AWS services. Price the path back before you move.
Cloud repatriation for teams under 50 engineers
What cloud repatriation takes for a team under 50 engineers: who replaces a failed disk, how to sort the bill, and why QA moves before production.
Choosing a SaaS hosting provider by who runs operations
Replay one outage across a PaaS, AWS or Azure with a consultant, and a managed private cloud to see whose engineer does each operating step.
Legacy application hosting inside the vendor's support matrix
The vendor's supported platform list decides where a legacy application runs. Three options as SQL Server 2016 and Windows Server 2016 age out.
What Business Support+ costs on a $40,000 monthly cloud bill
AWS Business Support+ costs $3,000 a month on a $40,000 single-account bill, and more across accounts or with a prepaid reservation.
Due diligence questions for a managed private cloud provider
Questions for a managed private cloud provider, grouped by the failure each one guards against, with the evidence a strong answer includes.
VMware renewal quote: options for a cluster of three hosts
Recount a Broadcom renewal quote for three ESXi hosts from the cores you run, and price the replacement hosts before choosing a VMware alternative.
AWS egress fees for a 10 TB month
Ten TB of AWS internet egress lists at $912.60. With NAT Gateway, cross-zone traffic and replication, one 10 TB month came to $1,349.07.
Before January 12, 2027: triaging Windows Server 2016 by role
Windows Server 2016 end of support lands January 12, 2027. Give each server role its own action and cutover weekend, and know where ESU applies.
Testing SQL Server 2016 upgrades on a restored copy
SQL Server 2016 end of support passed on July 14, 2026. Rerun the heaviest cycle on a restored copy running 2022 or 2025 before the upgrade.
Why does my Azure bill keep going up?
Sort an Azure bill increase into added work, leftover resources and rate changes in about an hour in Cost Management, then act on each one.
Test environment provisioning from approved configurations
What a QA configuration pins above the golden image, who approves each version, and the order Petrichor builds a stack from it, with a check after each layer.
High availability for a small SaaS backend
Which servers need a backup twin, how the database decides who's in charge, and how to time a failover drill.
Compatibility test matrices across operating systems and databases
Five operating systems and five databases make 25 cells, and vendor platform tables cut them to 21. Decide which run on every build, and where.
Reproduce a customer bug on their exact version
Collect the customer's patch levels in the first reply, keep busy releases built, and give each stack a known dataset and a test license.
Migration rehearsal: cutover timings and the rollback deadline
Time each cutover step against a restored copy of production, and fix the clock time for the rollback decision before cutover night.
Counting engineer hours and support fees in cloud TCO
Price one application's AWS or Azure cost with its support fee, open commitments and engineer hours, then find the quote a provider must beat.
Database restore tests on hosted SQL Server and PostgreSQL
Turn each database's recovery targets into backup settings, then time a restore that shows whether a host can meet them.
What a passing test proves in dev, QA and staging
Dev, QA and staging each answer a different question. Here's how to set up each one so its tests mean something.
Managed hosting responsibilities, assigned task by task
Before you pick a provider, write down who does each task, who signs off, and how you'll know it got done.
Private vs public cloud once the application is in production
Four checks for an application on AWS or Azure: the managed services it calls, the versions it pins, shared hardware behavior, and ticket owners.
Resetting a QA environment between test runs
Choose a reset method by the state it leaves untouched, and check the environment against a written starting state before the next run.
Release management when a hosting provider runs production
What a software vendor and its hosting provider write down before a release: who can stop it, the artifact digest, and the schema change.
Your app works. Is it ready for real customers?
Five questions to answer before real customers use an app you built with AI tools: where it runs, who can get in, backups, patching and cost.
How to read a cloud bill in ten minutes
A plain walk through a cloud invoice for business owners: the five lines that carry most of the cost, and the cost the invoice leaves out.
Resources
Company
About Petrichor
Trinsic Technologies has run its own private cloud since 2015. Builders, software teams and business owners get the environments they need from Petrichor, operated by the engineers who run the hardware.
Talk to a consultant
Tell us what you're stuck on. A consultant replies about what you're running and what we'd build.
Editorial policy
Who writes the guides and how to send a correction.
Website privacy notice
What this site records when you request a call or use the cost calculator.
Site directory