Friday, 7:30 PM. Client calls.#
"Hi, it's Smith. Listen, the living room lights stopped working. Again."
You know the feeling. You keep an SD card, a PSU, a spare Raspberry, an Ethernet cable, and three screwdrivers in your van - just in case. Half an hour ago you were at dinner. Now you're driving 2 hours across town to spend 10 seconds clicking Reload Config in the client's Home Assistant panel.
This is a typical Smart Home installer ticket. 8 out of 10 such tickets are things you'd fix in 5 minutes - if you only had access to the client's HA panel.
This article shows exactly:
- Which problem categories you close remotely (and how fast)
- Which require a truck roll, but you diagnose BEFORE leaving home
- How much an installer with 20 clients really saves
- How to set it up step by step
Distribution of service tickets - real data#
Based on stats from ~100 Smart Home installations (Home Assistant, Domoticz, OpenHAB) observed over 12 months, tickets roughly split like this:
| Problem category | % of tickets | Without remote access | With remote access |
|---|---|---|---|
| Broken automation / YAML | 40% | 2-hour drive | 5 min remote |
| Integration / add-on update | 25% | 2-hour drive | 10-15 min remote |
| Config reload / HA restart | 15% | 2-hour drive | 2 min remote |
| Dead add-on / container | 10% | 2-hour drive | 5-10 min remote |
| Hard crash HA / hardware failure | 10% | 2-hour drive | Drive + diagnosis |
The first 4 categories (90% of tickets) are purely software - all you need is access to the client's HA web panel. The 5th category (10%) requires a truck roll, BUT with remote access you know what happened before you get in the car.
Category 1: Broken automation (40% of tickets)#
Typical scenario#
Client replaced a bedroom device (new Zigbee bulb), the entity name in HA changed from light.bedroom to light.bedroom_2. The "turn off lights at 11 PM" automation still points to the old name. At 11 PM the bulb doesn't turn off. Client calls next morning.
How you fix it remotely#
- Open the client's HA panel via SmartHomeEntry:
https://smith.smarthomeentry.com - Settings → Automations & Scenes → click "Bedroom night" automation
- Change target from
light.bedroomtolight.bedroom_2 - Save → Run (test)
- Bulb turns off, client gets a text "done"
Time: 5 minutes. Cost: zero driving, zero fuel, zero ruined dinner.
What you WON'T do remotely#
If the client has a NEW device and it wasn't paired with HA - you must be physically present to put it in pairing mode (reset button on the device). But once it's paired - everything else is remote.
Category 2: Integration update (25% of tickets)#
Typical scenario#
HA auto-update installed a new ESPHome integration version. Turns out they changed breaking API - all 15 client sensors show "unavailable". Client sees red dots in the panel, panic.
How you fix it remotely#
- HA panel → Settings → System → Logs → find the error
- Google: "ESPHome 2026.4 breaking change" - find the GitHub issue
- Solution: add 3 lines to the
esphome:section of config - Open Studio Code Server add-on (through the same tunnel), edit YAML
- Reload ESPHome integration
- 15 sensors back online
Time: 10-15 minutes. Cost: zero driving.
Pro tip#
SmartHomeEntry Pro plan ($25/mo) has webhook alerts - configure HA to send warnings when an entity goes unavailable. You get a Slack alert BEFORE the client notices. Bonus: you look like a prophet when you call the client "hey, I know you have a problem, fixing now" - before they've even realized something's wrong.
Category 3: Config reload (15% of tickets)#
Typical scenario#
Client tried to add something in YAML themselves (thought it would be easy). Broke it. HA shows "Configuration invalid". No automations work. Panic, phone call: "NOTHING WORKS!"
How you fix it remotely#
- HA panel → Developer Tools → Check Configuration
- Error: "line 47 in automations.yaml - expected block sequence"
- Studio Code Server → open
automations.yamlline 47 - Remove the manually added incorrect indentation
- Check Configuration → Valid
- Reload Automations
Time: 2-5 minutes. Client thinks you're a wizard.
What you WON'T do remotely#
If the client broke HA config SO badly that HA won't even start - you need a shell to the OS (SSH to Raspberry). Then you either visit or ask them to launch the Terminal & SSH add-on (if it was enabled). SmartHomeEntry is a tunnel to the PANEL, not a shell to the OS.
Category 4: Dead add-on / container (10% of tickets)#
Typical scenario#
Node-RED add-on freezes after 4 weeks (typical thing - memory leak in some module). HA ↔ Node-RED hybrid automations stop working.
How you fix it remotely#
- HA panel → Settings → Add-ons → Node-RED → Restart
- Check Status after 30 seconds - Running
- Smoke test: trigger an automation using a Node-RED flow
- Works
Time: 5-10 minutes.
When you drive#
When restart doesn't help (add-on won't start, constant loop), and Supervisor itself starts having issues. Often points to a dying SD card or OOM. You drive - but you already know it's hardware, you bring a new Samsung Pro Endurance SD card.
Category 5: Hard crash HA / hardware failure (10% of tickets)#
Typical scenario#
Client calls: "nothing works, panel won't open". You check https://smith.smarthomeentry.com from your phone - no response. Check tunnel status in SmartHomeEntry dashboard - offline for 23 minutes.
Diagnosis BEFORE the drive#
This is the key value of remote access for this category. Even if you drive - you drive with specific knowledge:
| What you see | Probable cause | What you bring |
|---|---|---|
| Tunnel offline + ping to client router fails | Client router died or no power | Nothing - client restarts router |
| Tunnel offline + router ping OK | Raspberry died (SD, PSU) | SD card, PSU, backup image |
| Tunnel online, HA returns 502 | HA crashed (OOM, kernel panic) | Laptop + Ethernet cable for SSH |
| Tunnel online, HA 200 but entities unavailable | Zigbee coordinator died | New coordinator (Sonoff/ConBee) |
Without remote access - you drive blind, take everything "just in case", spend 30 minutes on diagnosis at the client's, then often drive back for the missing part.
With remote access - you drive with the right part and close the fix in 45 minutes instead of 3 hours.
ROI - real savings for an installer with 20 clients#
Assumptions:
- 20 clients × average 1 ticket/month = 20 tickets monthly
- Truck roll = 2h of work (1h driving + 1h at client)
- Problem distribution as per table above
Without remote access#
- 20 drives × 2h = 40 man-hours/month
- Fuel + vehicle depreciation: ~20 × $15 = $300/month
With remote access (SmartHomeEntry Business $60/mo)#
- 90% of problems (18) × avg 8 min remote = 2.4h
- 10% (2) × 2h drive (with diagnosis) = 4h
- Total: 6.4 hours/month
- Fuel: 2 × $15 = $30/month
Savings#
| Metric | Difference/mo | Difference/year |
|---|---|---|
| Work hours | -33.6 h | -403 h |
| Transport costs | -$270 | -$3,240 |
| Value of recovered time ($30/h) | +$1,008 | +$12,096 |
| SmartHomeEntry Business cost | -$60 | -$720 |
| Net savings | +$948 | +$11,376 |
But more importantly: those 403 hours per year can go to new clients instead of the same old ones. At $450 margin per installation and 10 man-hours per install = potentially 40 new installations per year = $18,000 of additional revenue.
How to set it up - 5 steps#
Step 1: Sign up for SmartHomeEntry#
Pro plan (5 installations / $25) if you're starting, Business (20 / $60) if you have a portfolio. See plans.
Step 2: Install the agent at the client's#
Before the visit, check the client's local address in the panel. Picking "Home Assistant" in onboarding fills in localhost:8123 for you - but a tunnel created any other way falls back to localhost:8080, and on that value the tunnel comes up while forwarding to a port where nothing listens, so the subdomain returns 502. If the client's HA runs on a non-standard port, set it here.
On your next service visit, add a 60-second setup. In the HA terminal (Terminal & SSH add-on or Docker exec):
curl -sSL https://api.smarthomeentry.com/api/install/CLIENT_TOKEN | sudo bash
The agent configures SSH keys, creates a systemd service, connects to the relay. 60 seconds, literally. The client won't even notice.
Step 3: Test access#
Back at the office, open https://client1.smarthomeentry.com - should work like being on the local network. Log in with your admin HA password (which the client gave you during the original install).
Step 4: Configure alerts#
In the SmartHomeEntry panel, set offline > 5 min alerts → send to your email or webhook to Slack / ticketing system (Jira, Zendesk, ServiceNow).
Step 5: Update the client contract#
Add a clause:
"SmartHomeEntry provides remote monitoring and diagnostics. In case of service issues, we aim to respond remotely within 30 minutes of report."
This is your marketing edge. You can legitimately offer SLA 99%+ with proof (PDF reports from SmartHomeEntry). A competitor without this tool can't offer that.
Installer FAQ#
Can the client see I have remote access?#
Yes, HA logs every login in History → Logbook. You can present this as a feature: "in the panel you see exactly when I log in - full transparency". Builds trust.
What if the client changes the HA password?#
Your access will stop working. That's why in the service contract it's worth having a clause "client commits to informing about changes to HA admin passwords". Alternatively create a dedicated service account [email protected] with admin rights - the client doesn't touch your credentials.
Is it legal GDPR-wise?#
Yes, if you have a signed Data Processing Agreement (DPA) with the client. SmartHomeEntry keeps infrastructure data in the EU (OVH Warsaw), which simplifies compliance. DPA template for installers available after signup.
What if the client has multiple sub-networks (VLANs)?#
The agent must be installed on a host that has access to HA (i.e. the Raspberry with HA). Traffic goes outbound from that host to our relay, so client's VLANs don't interfere - the tunnel works independently.
Can I manage multiple clients from one panel?#
Yes, that's the main use case. Pro plan = 5 installations, Business = 20, Enterprise = 100. One dashboard, all online statuses, alerts, uptime history, health scores. See panel demo.
Can I give access to my own technician?#
Business plan has user roles - add technician John, he sees only his assigned clients. You see everyone. Every action logged per user.
What's the uptime guarantee?#
Our SLA: 99.9% relay uptime (~43 minutes downtime per year). Exceeding = pro-rata compensation. Historically we maintain 99.97%+ (~15 minutes per year).
Summary#
No tool will make your work 100% remote - 10% of tickets will always require a truck roll, and that's fine. But 80% of software problems you close from the desk in 5-15 minutes, and the remaining 20% of hardware issues you drive to with specific knowledge instead of a trunk full of "just in case" gear.
Realistically for an installer with 20 clients that's 40 saved hours per month. You can convert that into new clients, your own free time, or higher margins in offers (clients pay more for "remote service in 30 min" than for "on-site visit within 48h").
Start with Pro Plan ($25/mo, 5 installations) - enough for first tests with 3-5 clients. Once you see the ROI, switch to Business ($60/mo, 20 installations). The first month usually pays back the subscription cost 10-20×.
