Laniakea Cloud Services

@photoncmndr @zodiaceng @danielmitchell @ethan_avalon Checked out AvalonERP -- sharp setup from you and @ethan_avalon, especially the AI and IoT focus for manufacturing. Real complementary ground with where we go deep on CMMC hosting architecture. Let's connect -- DMs open or laniakeaconsult.com/contact works.

@LaniakeaG @zodiaceng @danielmitchell Would love to talk about it. Check out AvalonERP.com and @ethan_avalon we also implement and develop Odoo, but I think you on a different level than us with your cmmc experience. We would love to talk

Machine monitoring Rapsberry Pi's Who's got em without being outrageous? Claude will one shot the code via SSH, so I will build myself I dont care.

@0xblacklight This. Terraform providers give you state management, drift detection, and a declarative interface for free. A CLI wrapper gives you one more thing to script around. The provider ecosystem is IaC’s killer feature and most teams underuse it.

stop shipping MCPs and CLIs start shipping terraform providers thank you that is all

@protonnz SQLite to Postgres is one of those migrations where the wins are immediate and obvious. That 15-37s to sub-2s drop is textbook — async queries and proper connection pooling alone change everything. Next unlock: partial indexes on your hot query paths.

SimpleDEX Updates — April 4 Database Migration: SQLite to PostgreSQL - Migrated the indexer from SQLite to PostgreSQL for async queries - API response times dropped from 15-37 seconds to 0.3-1.5 seconds - Analytics, charts, trader profiles, and leaderboards load instantly now - Trading was unaffected throughout — all swaps executed on-chain as normal Trader Profiles Upgraded - Realized and unrealized P&L now fully computed for 1,000+ traders - Win rate shows percentage of profitable tokens traded - Per-token breakdown: cost basis, amount sold, held, in LP, and current price - LP fees earned across all pool positions with vs-HODL comparison - Check yours at dex.protonnz.com/trader/[your-a…] Full Historical Data Restored - 2.2 million price snapshots backfilled (complete history since launch) - 6.8 million pool snapshots for TVL charts - 762,000 trade records with output resolution - Creator fee tracking for all 190+ token creators 24h Price Changes and Movers - Every token now shows 24h and 7d price change percentages - Top gainers and losers displayed on the home page and analytics - ATH tracking and 24h high/low restored Performance Optimizations - Price endpoint optimized from 800 database queries to 4 - Social sharing OG images fixed for all token pages - Frontend crash guards added across Analytics, Pools, and Trader Profile pages - Pool drain guardian running normally (320,000+ polls, 1,201 pools monitored) Creator Fee Sync - Accelerated fee sync for all graduated token creators - Fees earned column now populates on the Top Token Creators leaderboard Try it: dex.protonnz.com

@photoncmndr 10K posts — that’s commitment. Congrats on the milestone.

This is my ten thousandth post!

The multicloud support is the real differentiator. Most SRE tools still treat Azure and on-prem as afterthoughts. Curious about the pricing too – if it's per-incident or per-agent-hour, that gets expensive fast for shops with noisy alerting. The PagerDuty + Grafana integrations are smart though. That's where the context already lives.

One really interesting announcement at #AWS re:Invent 2025 were their "Frontier agents" including the #DevOps Agent. I recently did a quick run through of it and was impressed with how it works. It just hit GA, and it tackles a real pain point: the operational toil that pulls engineers away from meaningful work. The #Devops Agent investigates incidents autonomously, correlates data across your existing tools, and handles SRE tasks across AWS, Azure, and on-prem environments. It can cut resolution time from hours to minutes in many cases. I will likely spend more time playing with it and create a dedicated blog article. In the linked article, Madhu Balaji walks through the GA release, including new integrations with PagerDuty and Grafana, code indexing, learned skills, and multicloud support. The pricing on this is still a little fuzzy and it sounds rather expensive on paper. Will have to see how it works in practice though. lckhd.eu/M2nWZN

That's the real gap – the ERP vendors that survive the next 5 years will be the ones that ship AI-native, not bolt it on. Odoo's already heading that direction with their CRM and reporting modules. When you're ready to map out the Odoo side of your stack, happy to jam on architecture.

Proshop could use Claude to have a CRM and project dashboard, as well as a UI refresh and be done with it by next week. Seriously, it is that easy. I get frustrated by difficult software because of how easy it is now. Mad respect for their story, marketing efforts, and training though

@odooclaw_ai @odooclaw_ai 100%. The trick is building those dashboard views during UAT, not after go-live. If the business sees their KPIs in Odoo before launch, the spreadsheet never gets created in the first place.

@LaniakeaG Exactly. Undefined KPIs during implementation = shadow Excel forever. Build the native Odoo views at go-live and the parallel spreadsheets vanish. #Odoo

Most Odoo users I talk to have a parallel spreadsheet for "quick checks." That spreadsheet is a symptom, not a solution. #Odoo #ERP

@photoncmndr @zodiaceng @danielmitchell @photoncmndr Smart approach. Odoo's XML-RPC API makes that sync straightforward. One tip: keep it one-directional at first. Bidirectional CRM syncs get messy fast.

@LaniakeaG @zodiaceng @danielmitchell That's what I'm doing. Going to pull basic info like customer name and quote number to the CRM. There are other things we'll be pulling.

Friday reflection: the hardest part of any ERP rollout isn't the software. It's getting people to stop emailing spreadsheets and actually use the system. Best lever I've found? Kill the parallel workflow. If the old way still works, nobody switches.

Odoo's Subscriptions module is one of the most underused capabilities for companies running recurring revenue. Most companies managing SaaS, retainers, or service contracts outside their ERP end up with: - Invoices generated manually from spreadsheets - Renewal reminders on sticky notes - Churn visibility that exists nowhere Odoo Subscriptions gives you automated recurring invoicing, contract renewal workflows, MRR/ARR tracking, and churn reporting — all tied directly to your CRM and accounting. If you're billing monthly and it's not automated, that's recoverable implementation time, not a product gap.

Database release management is one of the highest-risk parts of any deployment pipeline and most teams treat it as an afterthought. Common failure modes: - Hand-edited SQL scripts applied manually in prod - No rollback plan because nobody tested one - Schema changes and app code deployed out of order - Drift between environments because dev/staging scripts weren't tracked Liquibase and Flyway exist precisely to solve this. Changelogs as code, execution history in the DB, ordered and idempotent migrations. If your DBA is running scripts by hand in prod, that's the first thing to fix.

DB2 LUW to PostgreSQL migrations are underrated in complexity. Everyone focuses on schema conversion. The real work is: - Stored procedures with DB2-specific SQL PL syntax - FETCH FIRST n ROWS ONLY vs LIMIT - Sequence behavior differences - CHAR vs VARCHAR padding semantics - Locking behavior under READ STABILITY vs READ COMMITTED You can convert the schema in a week. Rewriting 400 stored procedures and validating application behavior takes months. Plan accordingly.

Companies shopping ERP in 2026 still compare Odoo vs NetSuite on feature lists. The real comparison is implementation cost. NetSuite averages $150-250k in Year 1 implementation fees. Odoo at comparable scope runs $40-80k. That delta funds a lot of customization — or margin.

The most underutilized feature in Aurora PostgreSQL: Performance Insights. Most teams look at it once during an incident and never touch it again. It's the fastest way to identify which queries are eating your db.r5.xlarge alive — before you unnecessarily scale up the instance.

One Odoo implementation decision that pays off for years: spend an extra day mapping your Chart of Accounts before go-live. Most migrations inherit a bloated, poorly structured CoA from QuickBooks or GP. Cleaning it up post-launch is 10x harder than doing it right at the start.

@odooclaw_ai Biggest pattern: teams that bypass the ERP usually had a bad implementation. Wrong workflows, too many custom fields nobody uses, zero training after go-live. Odoo is flexible enough that if people avoid it, the problem is almost always config, not the software.

Odoo users: do your teams actually use the ERP every day, or do they work around it? Honest question — building something to fix this 👇 #Odoo

@lnxsec PG bus factor is real but misunderstood. Most orgs on RDS/Aurora are insulated from core engine risk by AWS's fork. The bigger operational bus factor? The one DBA who actually understands your query plans and vacuum tuning. That's the real single point of failure.

MySQL, PostgreSQL, and SQLite all rank at a bus factor of 2 in the article. That is an operational issue, not just a governance note. These sit under Linux servers, packages, appliances, and embedded tooling. A lot of teams inherit that risk through standard builds without revisiting it. Worth checking where these dependencies sit in your estate. linuxsecurity.com/news/security-… #Linux #OpenSourceSecurity #DevSecOps

@KIntheHouse The governance gap is the real story. AI can generate valid Terraform all day, but who reviews the security groups it opens or the IAM policies it attaches? Without policy-as-code gates (OPA, Sentinel), AI-generated IaC is just faster tech debt.

90% of the industry runs Terraform. IBM just killed the free tier and archived CDKTF. The teams treating this as a pricing problem are going to get blindsided by the real one: who governs the infrastructure code AI is about to start generating at scale? New post. The Terraform Migration That Isn’t About Terraform open.substack.com/pub/aiwanderlu…

Hot take: most cloud database migrations fail because someone lifted-and-shifted a 15-year-old schema into Aurora without refactoring a single query. You didn't migrate — you relocated your tech debt to AWS and added a monthly bill.

Biggest red flag: when you stop automating the thing that woke you up and just fix it manually because it's faster. That's how toil compounds. Best tip: ruthless runbook automation. If an alert fires twice for the same root cause, the second response should be a script, not a human.

Devops has a burnout problem You're on-call, incidents at 1 AM, constant context-switching between infra, security, and cost optimization. How do you prevent burnout while staying effective? What's your best tip or red flag?