Support plans have been determined. This decision is based upon providing competitive pricing and furthering development, while optimizing support channels for users.
First, support will always be free throughout the knowledge base and community via Reddit and Discord. I will pop in from time to time to assist when possible and assess where opportunities to improve the panel + KB exist. Nothing is ever perfect and certainly something at its inflection point, being its first public release, has much to grow upon. apnscp has been in use since 2002 and many of the features came about not from storyboarding but from customer interaction; "that's cool, but what about..." or a ticket that came in on a pressing issue. These tickets provide opportunities to automate flow, such as providing storage amnesty that offers users an automated process whenever they exceed to storage capacity to remedy the problem and keep their site running (quota => storage_boost in
Second, hosting is still a technical field. Protocols are built upon protocols; when its predecessor fails the whole pyramid comes crashing down. Some incredible people have done a phenomenal job making things easier, but we're still far from perfect; having companies2,3 that exist solely disinfect compromised sites are evidence we still have a way to go to ensure safe hosting.
apnscp is a highly opinionated approach to hosting that is designed to allow you to go to the beach without worry. In fact, it's allowed me to travel overseas without worry. Many of the monitoring idiosyncrasies in Argos were developed from my experience when a service was acting peculiar - e.g. Postfix memory usage swelling indicating an increase in smtp processes, which in turn suggests a high outflow of mail - or didn't recover as it should such as a large MySQL InnoDB storage pool that could take a few minutes to properly load, so checking for lock is necessary before attempting to restart.
Support is necessary when a product fails to deliver its service. Support also offers itself as the path to least resistance, which presents a threeway conundrum between raising prices to hire talent, providing lackluster inexpensive support, or focusing on taking care of majority of customers through product coverage while letting go a few who need a more expensive, managed hosting solution.
Reflecting on GoDaddy's devilish approach
GoDaddy made an interesting move in 2014 to remove ticketing in favor of phone/live chat, which is fixed on the capacity of its support team. More tickets create longer wait times. Filing a ticket has a marginal activation cost as it's possible to open a ticket with, "It doesn't work". This can create a noisy pingback effect with substantially higher support costs through context switching between the customer and support member when the requisite information to resolve an issue immediately is insufficient thus necessitating a ping-pong between both sides. If support is swamped with meaningless quips of "It doesn't work" or "I don't understand why I can't do this", these queues swell mandating one of many solutions:
- Cost increases across the board to hire personnel to meet demand
- Pareto rule applies here; the few drive up costs for the majority
- Support personnel are hired overseas often inexpensively and with minimal interest in the company. Consequently, there is no incentive for the support personnel to provide a proper response often resulting in confusion or dissatisfaction
- This was an endemic problem with shared hosting in the early 2010s that led to its dissolution
- Issues are ignored or handled "eventually" several weeks later because simply there are not enough individuals to handle the workload, but the price remains the same. This is common in companies that have experience growing pains from increased user demand but an inability to hire sufficient talent to meet demand
- Hosting is complex with asymmetric, non-transferable knowledge. Several generations of a ticket benefit neither side. Having an issue resolved on its initial communication is preferred to an nth generation resolution, through which each successive generation decays some information that can be integrated to improve support or product flow
- As a corollary: why was this ticket originally opened? Ease of support or a defect in the product? Delay creates decay. Worse yet, having a serious concern persist for days on end produces negative sentiment both ways when legitimate issues are scattered among less important issues that need additional dialogue to resolve
- In GoDaddy's case, both parties suffer in a sadomasochistic managerial overload. In order to get support, you'll have to wait at least several minutes and at most a few hours. Often times the user gets discouraged and looks things up on their own. Employees are constantly working on tickets, which doesn't resolve the problem of meaningless ticketing of "It doesn't work", and customers are constantly clinging onto hope that it really isn't a 4 hour hold time... and you will lose your place in line if you disconnect!
- Great solution to maintain costs amid burgeoning support demands, but bad for both parties. Incentivizes employee performance on ticket turnover speed. Disincentivizes on propriety. Lose your ticket, lose your spot
Providing free community support + a paid model aims to eliminate this problem without raising costs. Opportunities for an immediate resolution exist on common issues and I will periodically look into other issues raised within the community to see where information can be improved.
If you need more immediacy paid support options exist.
Paid support exists in two tiers with a variable pricing depending on minutes, much like taking your car to a mechanic, asking a lawyer for help, or even too - from my experience - opening a ticket with your data center asking them to replace a hard drive you have in a plastic box beneath the server, because you're 800 miles away and just received an alert a drive died; that's $125.
Both support tiers are backed by SLA to guarantee a response time within the given window. Moreover, any support request that is determined to be the fault of apnscp in incorrect provisioning, bad code, or oversight will be refunded 100% and you'll have a special place in my heart.
Support includes logging into your server, with your consent, to evaluate and resolve the problem. Every resolution will be backed with an impartial explanation of the event. If, on the callback you're not present, then you also have a one-time callback number to resume support; we'll do our best to accommodate your schedule. Once we are on the issue, we will work with you to resolve it immediately without entering your issue back in a queue where its meaning can get lost.
Low priority support
Low priority support is available Monday through Friday, 8 AM to 10 PM EST (-0500 GMT). All initiated requests will receive a callback within 24 hours during normal Monday-Friday business hours. Support is $49 for the first 30 minutes and $25 for each additional 30 minutes.
High priority support
High priority support is available 365 days a year at all hours of the day without blackouts. Support is guaranteed to turnaround and contact you within 1 hour of initiating a support request. The cost is $249 for the first 30 minutes and $125 for each additional 30 minutes.
Hosting ain't easy
Technology is based upon existing technology and to have a firm understanding of the present, you must know the past; to understand the past is to know its history. I've based this model on my observations over the last 16 years in the hosting industry. Finding a balance between paid and free support is crucial as the evolution towards decentralized VPS hosting continues to thrive.
VPS is not always the best solution for your needs, so be sure to evaluate your needs before making the decision!