Multiply templated tasks in one yaml or json, or in folders

Now we can create templated tasks such as:
```
{
"template-id": "generic_mean_alert",
"dbrps": [{"db": "telegraf", "rp": "autogen"}],
"vars": {
"measurement": {"type" : "string", "value" : "mem" },
"groups": {"type": "list", "value": [{"type":"star", "value":"*"}]},
"field": {"type" : "string", "value" : "used_percent" },
"warn": {"type" : "lambda", "value" : "\"mean\" > 80.0" },
"crit": {"type" : "lambda", "value" : "\"mean\" > 90.0" }
}
}
```
Files count is growing and growing.
It will be great to have an option for keeping multiply in one json file.
It will be something like:

```
[
{
"template-id": "generic_mean_alert",
"dbrps": [{"db": "telegraf", "rp": "autogen"}],
"vars": {
"measurement": {"type" : "string", "value" : "mem" },
"groups": {"type": "list", "value": [{"type":"star", "value":"*"}]},
"field": {"type" : "string", "value" : "used_percent" },
"warn": {"type" : "lambda", "value" : "\"mean\" > 80.0" },
"crit": {"type" : "lambda", "value" : "\"mean\" > 90.0" }
}
},
{
"template-id": "generic_mean_alert",
"dbrps": [{"db": "telegraf", "rp": "autogen"}],
"vars": {
"measurement": {"type" : "string", "value" : "mem" },
"groups": {"type": "list", "value": [{"type":"star", "value":"*"}]},
"field": {"type" : "string", "value" : "free_percent" },
"warn": {"type" : "lambda", "value" : "\"mean\" > 20.0" },
"crit": {"type" : "lambda", "value" : "\"mean\" > 10.0" }
}
}
]
```

Also importing from subfolders can be a solution of this problem. Now kapacitor imports only root level tasks.

No implementation files, tests, or entry points are named. Start by locating the code that imports root-level Kapacitor tasks and inspect how JSON and YAML task files are parsed. Define whether completion means supporting multiple task definitions per file, recursive subfolder imports, or both, then add coverage for the chosen behavior.

Written by the indexing model from the issue text.