xia_meta.health.Health

class xia_meta.health.Health(**kwargs)

Bases: EmbeddedDocument

Define the health situation

Mathematical Definition:

In the ideal situation, a health object is a series of downtime events. For example, for a one-month-period, your laptop is expected to fail: * Twice because of network issue, downtime of each will be 16 seconds * 1% chance to get you a disk failure(unlucky). Downtime might be 2-3 days for restoration

Each event is considered following Poisson Process and thus could be calculated for event combination. Suppose λ1t1, λ2t2 for event1 and event2. * event for event1 or event2 * event for event1 and event2 occurs at the same time The challenge is that the operation will become complex because the varies downtime introduced. That’s why we defined a standard downtime list frequency[0:27]: * Period is fixed at 2 ** 26 seconds * frequency[i] = downtime of 2^i seconds occurs how many times

For the above example: frequency[4] (16 seconds = 2 ** 4) = 2 * 0.0386 (1 month has 0.0386 * 2 ** 26 seconds) frequency[17] (2-3 days is about 2 ** 17 seconds) = 0.01 * 0.0386

Appendix: predefined overlap probability map

Pick a simple example:

• period = 8s, downtime_1 = 1s and downtime_2 = 2s

• two downtime could be placed at any moment in the period thus a point of the surface of 8*8

• The overlap means the abs(downtime_1 - downtime_2) < 1.5 (attention, the downtime time point is at the middle)

• So the result should be: 1-((64-1.5)/64)² = 4.632%.

__init__(**kwargs)

Methods

__init__(**kwargs)
from_db(**kwargs)	Transform database data to internal data
from_display(**kwargs)	Transform Display data to internal data
from_mtbf(hours, downtime)	Get Health object with Mean Time Before Failure
get_display_data([lazy])	Get Display data (visualized form of data)
get_error_rate([max_errors])	How much chance that max errors occurs
get_frequency(total_frequency, downtime)
get_raw_data()	Get Raw data (materialized form of data)
get_runtime_data()	Get Runtime data (calculated from raw data when runtime is activated)
get_sample()	Get a sample document data
set_new_downtime(downtime)	Set a new downtime in the case of specified situation
to_db([ignore_unknown, no_class])	Translate the runtime data into raw data (materialized form of data)
validate()	Validate if all of the component of document follows the predefined rules

Attributes

LENGTH
frequency
overlap_map
period

classmethod from_db(**kwargs)

Transform database data to internal data

Parameters

**kwargs (object) – python dict got from system

Returns

python dict object

classmethod from_display(**kwargs)

Transform Display data to internal data

Parameters

**kwargs (object) – display object got from front end

Returns

python dict object

classmethod from_mtbf(hours: int, downtime: float)

Get Health object with Mean Time Before Failure

Parameters

• hours – Mean Time Before Failure in Hours

• downtime – Downtime in seconds

get_display_data(lazy: bool = True)

Get Display data (visualized form of data)

Returns

python dict object

Notes

We will return the detail form if it is possible

get_error_rate(max_errors: int = 0) → float

How much chance that max errors occurs

Parameters

max_errors – max error number. Attention max supported error number = 170

Returns

probability in float[0, 1]

Return type

object

get_raw_data()

Get Raw data (materialized form of data)

Returns

python dict object

get_runtime_data()

Get Runtime data (calculated from raw data when runtime is activated)

Returns

python dict object

classmethod get_sample()

Get a sample document data

set_new_downtime(downtime: float)

Set a new downtime in the case of specified situation

Parameters

downtime – new downtime in seconds

to_db(ignore_unknown: bool = False, no_class: bool = False)

Translate the runtime data into raw data (materialized form of data)

Parameters

• no_class (bool) – Don’t save the class name into document

• ignore_unknown (bool) – When a field is not defined in the document, should we send it back to the database

Returns

python dict object

validate()

Validate if all of the component of document follows the predefined rules

Returns

Always None. No exception raised means the validation passed